Log in

View Full Version : Creative Assembly AI marketability



parcelt
06-30-2006, 13:00
This forum has recently seen a few excellent discussions on the AI in TW games. I've read these with great interest as for me -and it seems most people here- a major improvement of AI is more important then anyting else. Personally, I would be willing to go back to Shogun's graphics if that would get me even just modest improvements in tactical and strategic AI... not to mention diplomacy - on a side note, there are no excuses for RTW's horrible diplomacy system, please CA go and license some CIV IV code and be done with it (not meaning to rant, just some particular frustration as this is really not something that involves pushing the bounderies of today's technological possibilities, as may be the case for the tactical AI. It could easily have been done so much better).

However, according to what seems to be general consensus on the forum, cute graphics are just too important from a marketing point of view, unlike the AI, and therefore AI development will continue ranking second.

I fail to agree here, and the proposition I would like to make is this: why not drop the graphics and adopt superior AI as the game's main selling point? It is a strategy game, after all. I am not saying graphics should be ignored, just that priorities should be re-arranged ánd that this will not need to lead to a marketing fiasco.

I know a thing or two about marketing in the gaming industry and I am baffled by Sega's apparent fixation with graphics as the one and only selling point. If this were true, why would other platforms even exist next to the PC?

Although there may be different types of gamers (let's just say, those that want AI first and those that want graphics first), it is not a matter of who you will sell to. Us AI-gamers will buy the game anyway (even with current poor AI its really the only thing out there), the graphic-junkies will also buy it in any case because it'll look nice and shiny in a preview or demo anyway, also if the graphics budget is scaled down a bit.

The difference is that under the current state of affairs, the AI gamer ends up disappointed. Why not make it the graphics-junkie who's disappointed? I find it hard to believe that this will have a negative overall effect on sales (any lost sales on that end would be compensated by extra sales on by real strategic gamers).

I understand where it comes from, with the marketing people currently in charge coming from Sega and all..... still, this is a different market, and it might be worthwhile to re-evaluate priorities... what might it mean to have a strategy game acclaimed for new ground-breaking AI, instead of just another set of fancy pictures that'll be outdated in 6 months anyway?

Just something I would consider if I were on the team.

Mount Suribachi
06-30-2006, 13:07
not to mention diplomacy - on a side note, there are no excuses for RTW's horrible diplomacy system, please CA go and license some CIV IV code and be done with it (not meaning to rant, just some particular frustration as this is really not something that involves pushing the bounderies of today's technological possibilities, as may be the case for the tactical AI. It could easily have been done so much better).



Civ IV? hell, the diplomatic engine in Alpha Centauri is 10 times better than RTWs, and that game is 8 years old!

As for graphics over AI, its just part of an overall trend at CA to be less hardcore and more mainstream. Whether this decision has been taken by CA or their publishers (Activision or SEGA) we will probly never know.

Kambyses
06-30-2006, 14:12
Dear parcelt ,I think you said what many TW fans think.If they asks me ,I will say they use the same RTW Graphic (With only a few improvments) ,Because it is nice enough and myself still can't enjoy the game with full performance because of hardware issues:oops: .While AI needs many improvements ,Both in Battles and especially in Campaign map.I'm sure you've experienced this ,I offered to a rival faction which was nearly destroyed and had only one settlement and my powerful army was ready to ruin it to "become protectorate" ,But it refused even when I offered them one provence and 100000 Dinarii:dizzy2: ,Seems the AI cannot find the true way.

However ,I don't think they payattention to it.My experience showed that they will do what have decided.I wouldn't risk this time ,And If the AI be the same ,I will not pay money for it.

Regard
-Kambiz

A.Saturnus
06-30-2006, 22:19
Actually, I'm convinced that CA did not intentionally degrade the AI. I'm also convinced that during the development of Rome, more time was invested into AI than during that of MTW and also that the AI of Rome is actually more complex than that of MTW. What makes Rome's AI perform so bad is the increased number of variables that play a role. Units can behave in more different ways and have more special features. The AI cannot cope with all the options it has.
Just look at the strategic map. In MTW the AI would shove armies aimlessly around, but it didn't matter much because of the province system. In RTW the same behavior leads to the failure to cluster larger armies and subsequent piecemeal destruction of the AI forces by the player.

It is not so much a changed marketing strategy that leads to bad AI but the complexity of the game puts heavier burdens on it. So CA doesn't fail to improve the AI because they don't deem it important but because they can't. At least not without investments they consider not worthwhile.

doc_bean
06-30-2006, 23:12
I know a thing or two about marketing in the gaming industry and I am baffled by Sega's apparent fixation with graphics as the one and only selling point. If this were true, why would other platforms even exist next to the PC?


1. Console gamers and PC gamers are a different breed

Console gamers complain about how much a PS3 is going to cost, a machine that will last them about six years. PC gamers will spend twice as much money on a PC (if they're keeping it cheap even) and will upgrade about two or three times during this same period of time.

2. Graphics get you attention

It gets you in magazines, it gets you on gamespot, it gets you on tv. Free publicity !

3. Graphics might not sell, but bad graphics can hurt sales

People want to see their >1k rig preform at its maximum capacity, they expect good graphics. A lot will except decent graphics, but then the gameplay has to vastly superior to anything else out there, even then a lot of people simply won't buy it because it looks dated.

4. AI seems important, but most people won't notice

Valve is keeping statistics about episode one. There are people who haven't finished the game. It's about 4 hours long. Some people who bought CivIV probably haven't finished a single game yet. You can get more than 4 hours out of RTW before the crappy AI starts bugging you. A lot of people actually don't spend as much time playing the games they buy as you might think. Games have gotten shorter and shorter over the years, yet there's hardly anyone complaining, why ? Because a lot (most ?) people will only play a game for a couple of hours.

Kambyses
07-01-2006, 08:49
doc_bean ,Your talks is true ,But they are too generally.Your talk about Graphic is true but It's not cover all players.In fact for RTS (Real Time Strategy)gamer things are a few diffrent ,And TW series are RTS.Myself still playing "Axis &Allies" ,"Panzer General 2" and "Close Combat series" which are older than 2000.I'm not alone ,Many players Still playing "Close Combat 5"(Released 2000).It shows we as RTS gamers learnt Do not (Only) payattention to Graphic ,But we should consider the gameplay and AI quality too.

About games like "Civ IV" which you said few people have finished it ,Imo it is because of the "Game Play" Not The Graphic.

Regard
-Kambiz

doc_bean
07-01-2006, 13:23
doc_bean ,Your talks is true ,But they are too generally.Your talk about Graphic is true but It's not cover all players.In fact for RTS (Real Time Strategy)gamer things are a few diffrent ,And TW series are RTS.Myself still playing "Axis &Allies" ,"Panzer General 2" and "Close Combat series" which are older than 2000.I'm not alone ,Many players Still playing "Close Combat 5"(Released 2000).It shows we as RTS gamers learnt Do not (Only) payattention to Graphic ,But we should consider the gameplay and AI quality too.


Yes, even Starcraft still is played a lot these days. People will play old games even if they look outdated, they won't buy new games that look like those old games. Of course my statements are generalizations, but then, they're making games for masses, so that's the way they're probably thinking.




About games like "Civ IV" which you said few people have finished it ,Imo it is because of the "Game Play" Not The Graphic.



Hey, I liked CivIV :sweatdrop:

CivIV actually looks pretty good for its genre, what I was trying to say is that most people don't play games for that long, about 10hours would probably be a good estimate for how long the average person plays a single game. It might even be less. Research is needed ~D

Sabuti
07-02-2006, 01:27
I agree the AI is paramount. Recently a RPG game came out and there approached seemed to be to use the AI as a selling point over graphical improvements. It seems to me that graphics quaility can be matched or bettered easier than AI. If a TW copy-cat game came out with better AI, it could put a serious dent into CA. As a final note about graphics, the way I play battles, I'm fully zooned out anyway, so graphics improvements aren't that impressive to me. If I zoomed in to enjoy the beautiful intricate graphics, my left flank would crumble and my army would rout.

Dunhill
07-02-2006, 06:02
Game with better AIs are already on the market.

Take Command 2nd Manassas has a great tactical AI, and Birth of America and Conquest of the Agean have better strategic AIs. Of course these are from Indy wargame developers who don't have to pander to the lowest common denominator of the marketplace.

We have already seen a migration of player and modders to TC2M, and expect to see even more with the release of the strategic layer and multiplayer. The developers have been smart enough to allow the game to be very moddable and focused on making a great AI, the community cranks out unit sprites, scenarios, graphics mods and OOBs as required.

Cheers,

Ciaran
07-02-2006, 11:45
And how was it rated? I know an German magazine rated it 67% only, but that might not reflect international sentiment.

Mount Suribachi
07-02-2006, 19:56
Hardcore strategy games tend to get low review scores from gaming mags - they don't have the time or the inclination to spend hours reading the manual, playing the tutorials, getting to know the game mechanics etc etc

But that doesn't mean there isn't a market for them. EU2 and its derivatives consistantly scored 60-70% in gaming mags over here, yet its sold well across Europe. Reading reviews of the game its patently obvious the reviewers took one look at the game, thought "this is too hard", never bothered playing it, and then spend most of their one page review waffling about anything other the game, before ending with "if you like hardcore strategy games with lots of numbers you'll probably like this".

OldSchool
07-02-2006, 20:07
People who research a game and look for discussions of AI before buying:

Very few.

People who buy based on fanboy reviews written by a guy with limited exposure to a game that is advertised on his site:

Some, who now consider themselves informed.

People who buy based on top ten lists, best-selling lists, and advertising:

Quite a few, who at least have the morale bonus of being formed in deep ranks.

People who decide while standing in a Wal-Mart aisle, scratching their ass, and looking at the Pretty Boxes:

Most of them.

AI makes games fun to play. It makes them stay on your hard-drive a long time. It puts your job in jeapordy, because you OneMoreTurned all night. It doesn't, I'm sad to say, sell games.

(Just my two cents. No refunds.):laugh4:

Dunhill
07-03-2006, 08:41
The reviews are mixed, and its easy to guess which reviewers were grognards and which weren't

The fanboy reviewers point to the lack of 3D sprites first and mark it down. The grognards look at gameplay and attention to deatil and mark it up.

When the 3D sprite issue is brought up in the forums it is met with the overwhelming desire to trade better gameplay for the visual aspect which is already pretty good considering there are resolution selections and unit specific graphics (flags and uniform variety mix increases) that can chosen to push the game RAM up quite a bit for bells and whistles. I personally bought 2GB to have some of those bells and whistles. I think the 2D sprites do a hell of a job considering I think the game is much more immersive than TW. It's all about the movement rates and AI choices where the realism pulls you in. The filed of play is also much larger, and realistic, and road movement matters. Lots of little things all add up to a better gaming experience for me.

I don't get that with other games.

I must say that when STW first came out it was a great mix of strategic and tactical play that was quite enjoyable. I hope they can regain what's been lost. However, at the moment other games are doing better in each section individually. When they bring the pieces togethor they will have a winner, and trust me they are working hard on that.

Cheers,

sunsmountain
07-03-2006, 09:00
I know a thing or two about marketing in the gaming industry and I am baffled by Sega's apparent fixation with graphics as the one and only selling point. If this were true, why would other platforms even exist next to the PC?

Though I'm sure you think you know a thing or two, EA/Activision, and in this case, SEGA, have paid professionals who I would trust as a company a deal more in the area of marketing.

The real question is, why does everything have to be marketable?

Main point of these threads is that they at least alert CA to what their fanbase is really concerned about, namely AI. (and infantry running speeds)

doc_bean
07-03-2006, 11:01
Main point of these threads is that they at least alert CA to what their fanbase is really concerned about, namely AI. (and infantry running speeds)

The Org does not even represent 1% of the people who bought RTW, we're not that important to them :no:

Captain Fishpants
07-03-2006, 12:05
The Org does not even represent 1% of the people who bought RTW, we're not that important to them :no:

What a despairing - and quite frankly wrong - statement. Of course the long-term and expert players are important.

Before anyone jumps on me for not talking about M2TW, I must point out that I'm not part of the team working on that title. My involvement has been limited to one job in one very specific section of the game and I'm not in a position to comment in an informed way.

Peasant Phill
07-03-2006, 14:54
Captain Fishpants, thanks again for joining in our discussions.

Of course we (I pretend I speak for every .Orger here) know that CA considers long-term and expert players important. Proof for this is CA hirering a long-term/expert player to help in the development or fine-tuning of M2TW and you and other CA employees visiting us here at the .Org.

However, doc-bean has a point that marketingwise we the so called hardcore gamers are unimportant. We are small in numbers, compared to the mass of gamers that buy the game, so we only represent a small portion of the turnover and profit. And profit is still the reason why CA makes computer games.

Alexander the Pretty Good
07-03-2006, 17:52
Would be nice if somebody who was working on MTW2 popped on over here once or twice.

BeeSting
07-03-2006, 23:21
Captain Fishpants, thanks again for joining in our discussions.

Of course we (I pretend I speak for every .Orger here) know that CA considers long-term and expert players important. Proof for this is CA hirering a long-term/expert player to help in the development or fine-tuning of M2TW and you and other CA employees visiting us here at the .Org.

However, doc-bean has a point that marketingwise we the so called hardcore gamers are unimportant. We are small in numbers, compared to the mass of gamers that buy the game, so we only represent a small portion of the turnover and profit. And profit is still the reason why CA makes computer games.

I think CA developers are strategy game enthusiast like ourselves and make games that they themselves like to play. My hats are off for CA in coming up with a genre that combines the eye catching graphics with depth of real liife battle conditions. Need i mention their huge past success in keeping to this? And what they have done with RTW is revolutionary raising the bar of this form of strategy games out there from 2d to 3d, hence requiring highly sophisticated AI. It's only a matter of time after so many testing, trials and errors, with constructive criticisms and feed back from fans like us before they fine tune the AI that will give us some serious challenge, depending on the difficulty level of course. Regardless, for now, they have come a long way from STW and have me sold in anticipation for MTW2, while still playing RTW--thanks to their openness for allowing mods. I hate to say this but, you grossly underestimate the masses by saying that they are more interested in eye candies. I was an uneducated mass oblivious to tactics, the reason for flanking and so on till I came across STW. Their games although historically not 100% accurate has made many like me dig into history books.

Puzz3D
07-04-2006, 00:39
Regardless, for now, they have come a long way from STW and have me sold in anticipation for MTW2, while still playing RTW--thanks to their openness for allowing mods.
They have lost the handle on MP which is markedly inferior to STW MP. BTW, suicide generals were fixed in MTW/VI, but they are back in RTW/BI. Also, the RTW/BI battle engine is missing important features that were in the MTW/VI battle engine.

I wonder how much money I'm going to have to spend on total war games before the suicide generals are fixed? So far, I've spent $40 + $25 = $65 usd, and two years have gone by without it being fixed. I also don't get any use out of the MP part of RTW/BI because of the direction they have taken MP, and it doesn't look like this will change with M2TW.

Pras the Reaper
07-04-2006, 03:56
There seems to be a bit of a misconception amongst many forumites (and I do mean all the forums, not just the org) that graphics and AI are mutually exclusive. Simple fact of the matter is that we have graphics programmers and artists to concentrate on graphics and AI programmers to do the AI.

IMHO, the simple reason that people believe all the marketing attention has been focused on "eye-candy" is that currently it's the only thing that we can actually prove that they're improved. We've actually stated a number of improvements in graphics, animation, gameplay and AI both in interview and on the forums. Screenshots can highlight the graphical improvements, movies show off our new animations but there is no way for us to show improved gameplay and AI until the game is released and players can actually see it for themselves.

Duke John
07-04-2006, 08:05
Yes, you can, battle reports! There are plenty of fans who have played miniature games and who read battle reports. There are plenty who have Osprey books with their excellent blocks and arrows diagrams. People love them!

Plus it will give you good insight at what parts of the AI need improvement as you probably need to write about AI movement that is too embarrassing to write about.



I also don't get any use out of the MP part of RTW/BI because of the direction they have taken MP, and it doesn't look like this will change with M2TW.
The Lordz are having quite alot of fun testing NTW2 online. I was quite hesitant after all the GS problems, but it seems to work fairly well. All the OOS we got were caused by ourselves and there was relatively little lag and a single complete halt of the game. Of course you still need to ignore the public chat with people talking about doing things to another persons' mom.

sunsmountain
07-04-2006, 09:44
Perhaps I can share something, Duke John. It's 2 posts by JeromeGrasdyke, one if not the Lead Programmer of the team that created Rome:TW. I'm guessing he's very busy right now, so after hearing nothing for 1 month, and given the added importance of this thread because of not 1, but 2 CA members entering the discussion, I've decided to take the liberty of sharing his insightful comments. Here they are:

My original post to Jerome and hatcat (another programmer)


Originally Posted by sunsmountain

Dear gentlemen,

I know how busy you are, yet as a fan this question kept haunting me since release:

AI

Now in previews & diaries CA spoke of different levels of AI, and how they obeyed Sun Tzu's Art of War. I guess my main question is that I would like an insight into the battlefield AI.

In this topic
https://forums.totalwar.org/vb/showthread.php?t=52341

i'm having a pretty comprehensive discussion with Puzz3D about the current AI, but both of us don't really know the finer details of what we're discussing, making it difficult to formulate the problem.

Barring Imperial Glory and other copy-cats luring, I would love any possible sharing on this issue.

Sincerely,

sunsmountain


In reply, by Jerome ~:) (hatcat didn't bother :inquisitive: )


Originally Posted by JeromeGrasdyke, 17 August 2005

I've had a look at the thread ;) There's a lot of guessing and double-guessing going on there, and i'm not going to go into specifics on how the details of the Rome AI work. For one thing, it would take way too long! It's not a trivial amount of code by any means.

But in general, you are correct and the methods used in Rome were very similar to what was in Medieval, with some adaptions for the new game features. It was largely rewritten along the way for various reasons to do with the new system. The campaign AI was entirely new code, however.

Take it easy,

Jerome


And in reply to:


Originally Posted by sunsmountain

Cheers Jerome, but let me get this straight: AI is programmed "along the way"? Would it not deserve a month in its own right (doing nothing else - testing and testing until those spears/cav/swords/flanking works? Or does SEGA/Activision not allow this?

As for the complexity, it's certainly not trivial code i know, and without the full picture i probably cannot help you, but what i meant was: What is holding you back in creating the most powerful, incredible AI? (if, as a player, you dare set the difficulty to Very Hard)

Relaxed regards,

suns


The reply that is essentially a FAQ in its own right ~:)


Originally Posted by JeromeGrasdyke, 01 September 2005

Hehe, "along the way" was just a turn of phrase. In fact both the battle AI and the campaign AI had dedicated programmers assigned to them for over a year, who were assisted at times by other members of the team.

Essentially, the AI is an approximation of how the AI designer plays the game, refined in stages as he finds strategies which work or not. Some things cannot be implemented on a computer - for example, you say you just "see" that one given situation is better than another one. In order to write the corresponding code for the AI you have to know exactly why, and then you have to express it consistently and accurately in terms of numbers. That by itself makes it hard enough.

But then also consider that the game evolves underneath it as the AI systems are developed. Say the designers add a new feature to the game - then the AI programmer has to go back over his old code, which may involve a lot of linked strategies about how say Agents are used in combination with armies, and weed out what does not work anymore, and redevelop strategies for the cases which lead to exploitable gaps in the AI playstyle.

That unfortunately means that you can only write really good AI towards the end of the game's development, when the designers are making only minor changes and you've had the chance to actually play the game solo for at least a month or so... but by then you're inevitably under time pressure because the release dates are looming.

So in an ideal world, the game design would be completely finished and tested before the process of writing AI starts... unfortunately a real-world development studio does not work like that, there's lots of creative people always coming up with really cool ideas that just have to go in before the Alpha deadline ;) And maybe that's a good thing, because often the stuff that is added at the end is important balance and polish which makes the game much better.

We're always looking for ways to improve this process, and for BI and future projects from the Rome codebase the AI can be substantially improved, since in sequels you have that more-or-less fixed gameplay dynamic already. Hopefully that answers your question...

TB666
07-04-2006, 10:03
Yes, you can, battle reports! There are plenty of fans who have played miniature games and who read battle reports. There are plenty who have Osprey books with their excellent blocks and arrows diagrams. People love them!

Well Wikiman did write a battle report on the .com not to long ago about a mp battle they had which was fun to read indeed.
However not many people seemed to have noticed it :gossip:

Also I guess this is a battle report in the works
http://p223.ezboard.com/fshoguntotalwarfrm55.showMessage?topicID=2602.topic


Edit:
Found the battle report

Well Will, myself and the AI were defending a three ring settlement. The enemy was Kev playing two machines and Kim.

We lurched out of the East gate and cleaned up Scotland with a combined cavalry rush I then pulled mine back as a hole was busted in the western side. Will routed the whole Scottish army then flanked the southern army trying to make a run for the door. Not sure what happened to him after that as I was too busy! I plugged a hole in the western wall with one hard unit which I had pumped up to full experience, defence and weapons. This plugged the small hole they had made while my cavalry ran from the other side of the map. THis one little unit held their entire army back as they couldn't budge it. It was wedged in tight.

Meanwhile more and more holes were opening up all over the place. Will and I started making a coordinated withdraw to the second ring. Leaving some troops to die horribly for the good of the rest. I got virtually all of my archers back to the middle ring along with about 50 cavalry. Will, passing through the centre of town got massacred as he walked into a dead end as the defensive AI was passing along a road blocking his path with a mass of troops. His enemy caught up to him and cut him down from behind. Caught between friendly units and the enemy.

The western enemy army now started assembling for a heavy push top take the middle ring. The western army was virtually untouched and the only foot troops I had were archers and a unit of swordsmen. They started weakening all of the walls along the middle ring, hoping to rush me from all sides. The vastly outnumbered me. I made the executive decision to leave the AI to defend the middle ring (hehe), retreating to the inner ring as it had fewer access points. I pulled back my archers and cavalry and held tight.

The middle rings walls fell and Kim’s units poured in. You can guess the rest. :( I was munched.

-wikiman

parcelt
07-04-2006, 11:52
Thanks to Pras for joining in. You may be right in that it all comes down to a misconception, I hope it is but I am not entirely sure. It may be helpful to share the line of reasoning that led me personally to conclude that AI development is not the nr. 1 priority (although I do not think its mutually exclusive with graphics). I think most of this applies to a lot of the gamers on this forum, although I can obviously not speak for anyone else:

1. The failings of RTW (just to keep things in perspective, let me state that overall I feel RTW is a great game, before I focus on the negatives here).
- the things that went wrong in previous TW games (and were sometimes even fixed), yet went wrong again in RTW. Suicide generals, anyone?
- the things that really ought to be done right in any game coming out today that takes itself seriously, yet went wrong in RTW. Diplomacy is a pretty obvious example.
- how some very common compliants seem to be ignored, at least in the sense that they are not adressed in patches/expansions. Infantry running speeds and kill rates have been discussed over and over on the forums. I am no programmer but is it really that difficult to solve this by e.g. an Arcade mode option, a separate speed option, make it (more easily) moddable, etc?

All of the above gives a somewhat 'sloppy' impression. Now, I have a very high regard of CA and their staff, they have proven over again since Shogun that they know how to make a good strategy game. So then, why are these unnecessary mistakes / omissions here?

--> To me, images come to mind of a lone overworked AI programmer somwhere deep down in CA's basement, lacking resources (because those were allocated to e.g. graphics?) and time (because why spend a lot of time on AI development if the graphics are already done and we can sell the game with that?). The poor guy does what he can, but he can only do so much and is forced to make though choices, knowingly leaving things out he knows should be in, or leaving things in he knows could be done much, much better. It's tearing at him possibly more than it will later frustrate us gamers, but alas, this is how things are.... :shame:

Am I exagerating? Perhaps, but then you tell me why the above (and there are other examples) unnecessary issues exist.

2. All the demos, teasers, interviews, and other types of previews for both RTW and its expansions and MTW2, have an overwhelming focus on graphics/new 'spectacular' units/etc. True, these may be the things better suited for display in a preview, but to a lot of people the lack of attention to AI and similar issues (these could be discussed much more extensively in interviews/FAQs) is worrying. I refer also to another, excellent, thread on this forum, the title is obvious.

1 and 2 above combined are what makes me wonder, and worry, about how seriously AI development is really being taken. By 'seriously' I am not referring to programmers' view on these things or their capabilities (I am confident there), but rather to the amount of resources in terms of $$ and time being dedicated to AI development. Especially the comment regarding how AI programming is something taking place during the final phases of the game's production (this was new to me) is worrying in this respect.

Just my 2 cents. Thanks CA for being on the forum. And to that guy in the basement: keep it up, we're all counting on you!

Grifman
07-04-2006, 14:21
Game with better AIs are already on the market.

Take Command 2nd Manassas has a great tactical AI, and Birth of America and Conquest of the Agean have better strategic AIs. Of course these are from Indy wargame developers who don't have to pander to the lowest common denominator of the marketplace.

But as you note, each of these games only has half the AI requirements of the TW series, which has to have both a strategic and a tactical AI. As for Manassas tactical AI, you don't have multiple maps and terrain to deal with, and I suspect there are just a few unit type, infantry, cavalry, artillery. And BoA doesn't have near as many factions to play (each with their own unique map positions) as the TW series. Nor is there the economic/technlogical development of the TW series in these games.

You're really not comparing the same things here. Each of these games is much simpler than the TW series. Your comparison is simplistic and flawed.

Duke John
07-04-2006, 14:34
True, but the Take Command AI does need to take into account lines of communication, true line of sight, chain of command, use of roads and all of that on a very large map. It seems to me that developing an AI for R:TW was too much for CA to chew.

sunsmountain
07-04-2006, 17:03
Am I exagerating? Perhaps, but then you tell me why the above (and there are other examples) unnecessary issues exist.


Now i know Jerome is not part of this discussion, but did you read the comments above he made in private messages? Shall i quote for you? Here it is:


That unfortunately means that you can only write really good AI towards the end of the game's development, when the designers are making only minor changes and you've had the chance to actually play the game solo for at least a month or so... but by then you're inevitably under time pressure because the release dates are looming.

And to be fair, it's quite hard to write an exception to unit AI level code, specifically for generals, when such units do not carry their own tag. They don't carry their own tag because the game wasn't designed that way, and you can't go back and just add a tag.

And to be fair again, the diplomacy AI in RomeTW is not THAT bad, at least most things work, even protectorates. Sure the AI will agree or disagree to the best & the worst offers, but that only needs tweaking. Exactly the same as with battlemap AI.

As for infantry running speeds and kill speeds, cheer up! Those things CAN be modded, it's just not that easy (and i'm still searching for a non-invasive mod for those, too).

parcelt
07-04-2006, 18:20
Now i know Jerome is not part of this discussion, but did you read the comments above he made in private messages? Shall i quote for you? Here it is:


Quote:
That unfortunately means that you can only write really good AI towards the end of the game's development, when the designers are making only minor changes and you've had the chance to actually play the game solo for at least a month or so... but by then you're inevitably under time pressure because the release dates are looming.


Sunsmountain, not sure what your beef is, but your tone is rather unpleasant. Unnecessarily so, I would say, as we only seem to agree. I did read Jerome's comments and one of the points I was trying to make is that unduly time pressure in the last phase of production may be (in part) cause of some of those issues I described. And would not adding more resources to AI development alleviate that pressure somewhat?



And to be fair, it's quite hard to write an exception to unit AI level code, specifically for generals, when such units do not carry their own tag. They don't carry their own tag because the game wasn't designed that way, and you can't go back and just add a tag.

And how 'bout designing the game that way? As was possible for RTW after the suicide general-issue was identified in STW and MTW, and (at least partly) fixed in VI?



And to be fair again, the diplomacy AI in RomeTW is not THAT bad, at least most things work, even protectorates. Sure the AI will agree or disagree to the best & the worst offers, but that only needs tweaking. Exactly the same as with battlemap AI.

Now I have to ask.... is CA paying you to make these kind of statements?? I am only kidding, but honestly, you can't really believe this yourself. After two handfuls of patches and a zillion mods out there, these issues have still not been 'tweaked' to full satisfaction. It can't be done as it concerns inherent flaws in the game's design (above you noted yourself how rigid this can make things). I will not even touch upon diplomacy.



As for infantry running speeds and kill speeds, cheer up! Those things CAN be modded, it's just not that easy (and i'm still searching for a non-invasive mod for those, too).

Well, did YOU read what I wrote? I said I'd like an Arcade type of option to deal with this, or a more easy way to modd. That it can be modded by investing great amounts of time and effort is not the answer.


Just seems to me you're very eager to defend even RTW's obvious weakspots. Why not highlight them so they can be analyzed and improved?
Did you read this part:

(just to keep things in perspective, let me state that overall I feel RTW is a great game, before I focus on the negatives here).

I'm as big a fan of the TW series as the next guy. Just hoping the next one will be even better!!

OldSchool
07-04-2006, 19:10
I've been following these AI threads with some interest. I would probably follow them with more interest if they didn't automatically turn into complaint threads. Anyway, interesting stuff combined with some ridiculous statements such as "AI is easy", etc.

Programming AI is not easy. Nothing about it is easy. I can't think of one single game that I would consider to have outstanding AI. The path-finding difference between the RTW strategy map and the old Risk-style maps is huge. The difference with Civ-type games is also huge. I assume, in a completely uninformed way, that the Civ games use a shortest-distance-between-two-points method combined with a check of nearby squares for movement point bonuses/penalties. RTW appears (again uninformed) to use the same basic principal with the added headache of impassable terrain, ford points at rivers, etc. I would imagine that right now there is a guy at CA cussing at his keyboard, throwing things at his monitor, trying to work out just how to get the strategy map path-finding just right. I would also imagine that if you walked in and said "Hey bud, it's easy. Just do it like gamexyz.", that you would not walk out alive. It's one thing to criticize, which is natural and legitimate. It's another thing to assume that it's easy.
Sitting back, coming up with a logical solution to artificial decision making is easy. Actually implementing that solution, taking into account all the variables and factors (some of which have nothing to do with the logic of the solution, such as cpu usage, available animations, budget, schedule) is pretty damned hard.

Since this is in a MTW2 forum, I should say that I am optimistic that the strategy map AI will be much improved in that game, and for me it is a welcome improvement over the old Risk-style maps. The changes to the battles are not so welcome, but, as that has been well-complained about, I don't see the need to add mine here.

As far as the bashing goes. That, I think, is the result of having been spoiled. So RTW didn't have the magic you've come to expect from a TW game. So what? Get over it. It was still quite fun to play (for me anyway). I've got an old hard-drive that I bootstrap into all my computers that has my "magic" games on it. The ones I keep coming back to long after I should have let them go. Right there next to Quake and Alpha Centauri and NWN sits Medieval Total War. I'd say I've got my money's worth from CA. I don't expect every game in the series to make it to that hard-drive, but maybe that's just me.

Man, that was pretty long and rambly and preachy for a noob post, wasn't it? Sorry 'bout that.

Puzz3D
07-04-2006, 19:32
Now i know Jerome is not part of this discussion, but did you read the comments above he made in private messages?
Yes, and it confirms my impression that they aren't going to fix most of this stuff in M2TW. They hardcode things that should be adjustable. Why isn't there a tag on the general's unit? It certainly needs one.

Also, mods can't fix most of the AI issues nor can they put features in the battle engine that aren't there. Even a simple thing like charge, run and walk speeds was coded in such a way that you can't change them independently. Movement speeds are extremely important to the gameplay. A mistake was made in the ratio of the walking to the running speed, but CA said they won't change the movement speed because the animations would have to be redone to prevent the units slipping or sliding.

Movement speed is basic to the gameplay, and you can't just go changing it with the terrain settings because that alters the effect of fatigue and the effectiveness of ranged fire. If you notice, fatigue rate is something you can't mod.

So it certainly appears that the game is being marketed to people who don't care about these things. If that wasn't true, RTW would have had a slower speed setting for all those customers who liked the speed of MTW/VI. I was told directly by a CA developer that they don't see anything wrong with the movement speeds. Incredible Sidekick also posted here at the org that the speeds were motion captured so are correct. He didn't say what they motion captured to get these speeds.

Divinus Arma
07-04-2006, 20:24
I think it is important to point out the following:


We were told in this thread that in an ideal world, the game design would be finished first, leaving the programmers to work on the AI without complication.

The community has been recently told that the game design is essentially finished. The release date is a few months away. What will they be doing for the next few months?


I find this promising.

sunsmountain
07-05-2006, 09:41
Sunsmountain, not sure what your beef is, but your tone is rather unpleasant. Unnecessarily so, I would say, as we only seem to agree. I did read Jerome's comments and one of the points I was trying to make is that unduly time pressure in the last phase of production may be (in part) cause of some of those issues I described. And would not adding more resources to AI development alleviate that pressure somewhat?

parceIt, no offense, but i did feel you were exagerating there and responded to balance things out. More resources is of course good for any game, but you only have so much.


Just seems to me you're very eager to defend even RTW's obvious weakspots. Why not highlight them so they can be analyzed and improved?

Both sides of the story need to be heard, else there will be no dialogue. I hear and see the negatives too often. CA will change nothing unless they see our point of view, which they certainly won't if we deny to at least consider or know their point of view. It's their game after all. They designed it this way for a reason, you can disagree, but if you don't want to know their reason for designing it this way, why should they listen to your reason for designing it that way?


After two handfuls of patches and a zillion mods out there, these issues have still not been 'tweaked' to full satisfaction. It can't be done as it concerns inherent flaws in the game's design (above you noted yourself how rigid this can make things). I will not even touch upon diplomacy

I was talking about the tweaking a programmer can do, not necessarily a modder. Do you know about inherent flaws in the game's design? I'm not asking you to name the effects of this, i'm asking you for the causes. See, you can say that the design is bad but you know nothing about the design. How can you give CA constructive feedback then? Just because those barbarians didn't accept your Ceasefire + 10000 gold + map information + etc., doesn't mean they designed it wrongly...

I am still dead certain that a better weighing of offers by the AI, and a better taking into account of relative army strengths, can fix most issues in diplomacy. I don't know the issues you're having, but as of patch 1.5, the number of inconsistencies in the diplomacy AI have dropped, in my experience. It's still hard to get a peace with a faction you have been at war with for quite a while, even if you decimate them, unless they offer it of course. But perhaps that's only logical, and asking for "full satisfaction" is perhaps a little too much.

Declaring war by the AI also seems random to some, but is actually quite similar to the match ups the MTW AI made, and therefore predictable. If you defend your cities (more specifically: regions) with large armies, the large armies of the AI will not in general move into those regions. If you also have trade rights or alliance it helps, but the AI will still attack you if you defend your regions poorly. If you block the path to your city with a single unit, a path finding bug (which is still there, i agree) will almost always force the AI to move away from your cities, as it tries to reach your city via the other, longer way around.

The decision to go the other way around is perhaps an "inherent flaw in the game's design", but i'm sure they will fix it in MTW2. Well, I hope so...


They hardcode things that should be adjustable.

In a team of programmers, you have to make some ground assumptions so you can get to work. Now that the work is done, you can go back to those hard coded parts of the program and change them. But know full well that for each and every parameter you change, all the other numbers in your program can suddenly start to become buggy, resulting in crashes etc.


Movement speeds are extremely important to the gameplay. A mistake was made in the ratio of the walking to the running speed, but CA said they won't change the movement speed because the animations would have to be redone to prevent the units slipping or sliding.

See, I didn't know this but it does explain why the ratio will stay that way. Nobody wants slipping and sliding (would be funny though).


If you notice, fatigue rate is something you can't mod.
They obviously thought that, if these things were right, they wouldn't have to be modded. They got the balance right in the past, perhaps they became overconfident. Or, given the initial motion captures, the other stuff like fatigue and ranged fire parameters simply followed from that.

But we can't ask CA to do all the motion captures all over again, that would cost too much money while only catering to the few. What can we ask them to fix this? Perhaps change the terrain modifiers for cavalry and infantry separately, would that solve it? Or would the animations look really dodgy?

HarunTaiwan
07-05-2006, 12:33
It's really hard to make an automobile. It's really hard to make a printing press with 10,000 moving parts. It's really hard to paint a masterpiece. It's really hard to create a vaccine. It's really hard to write a good novel.

and I am sure it's really hard to make a computer game with good AI. But, that's CA's job and why they earn the big bucks.

Now let's say we sat down and wrote down on paper what rules we would want a computer to follow when fighting a medieval or ancient battle. What's on top of your list?

Don't suicide your general unit.
Don't melee attack with missiles.
Keep missile units behind spear units.
Attack in formation (i.e. in a battle line not unit by unit.)

Now, maybe it is truly hard to program this, so be it. (In Madden football does the quarterback perform QB Draws for most of the game?) If it is so hard, why do the game companies keep jumping the gun by creating Campaign Maps that they KNOW the AI can't handle. (Though, admittedly players can make good use of it.)

In fact, CA even shipped RTW with the load-save bug, which admits they think their AI is so poor they don't even need to let it have a memory - hell, just let the AI work from scratch each turn! Okay, that's harsh - the alternative explanation is that they shipped RTW with a major AI bug that they didn't even know existed. (ouch.)

However, I think we can all admit that many game companies, not just CA, focus on graphics, graphics, graphics. Pointless Waste of Time had an excellent essay on this point. Some company was extolling their graphics to the point of being able to see the sweat roll down the players' arms. ooooooh. This means that either all game companies are dumb or that, yes, graphics sells enough games to make it profitable to improve them.

What's weird to me is not working on MP, which by a stroke let's players play against superior AI without needing any actual AI programming.

HarunTaiwan
07-05-2006, 12:41
But we can't ask CA to do all the motion captures all over again, that would cost too much money while only catering to the few.

That's a good one. I will tell that to my customer the next time they require me to re-work an entire order of product because it doesn't work properly.

BTW, software is the ONE industry where replacing faulty product relatively cheaply is possible.

TB666
07-05-2006, 12:49
In fact, CA even shipped RTW with the load-save bug, which admits they think their AI is so poor they don't even need to let it have a memory - hell, just let the AI work from scratch each turn! Okay, that's harsh - the alternative explanation is that they shipped RTW with a major AI bug that they didn't even know existed. (ouch.)

The load/save was what you say it was, a bug. A bug that took around 6 months of thousands of players playing this game alot to notice.
The AI did have a memory, it just wasn't working properly.
Ever since the bug was discovered I only had 4 seiges broken total, the rest of the seiges during the pre 1.3 patch worked fine.
I doubt CA shipped RTW knowing the bug existed.

Kralizec
07-05-2006, 13:07
But denying the load/save issue exists alltogether, and claiming that it's actually a "feature" allowing the AI to reconsider its moves? :inquisitive:

Not exactly class customer service.

TB666
07-05-2006, 13:20
But denying the load/save issue exists alltogether, and claiming that it's actually a "feature" allowing the AI to reconsider its moves? :inquisitive:

Not exactly class customer service.
And I wonder if Activision wasn't behind that because as soon as they moved to SEGA they suddenly confirmed it and fixed it.
The same with Vampire: Bloodlines(also Activision) where the devs responded to the reports and proof of huge memory leaks and unoptimised engine by simply that the people's computers was under the minimum specs while ignoring the fact that the people's computers was way over the recommended specs.
Also the "feature" bit was a misunderstanding I think. CA wasn't trying to pass it off as a feature but merely trying to explain how the AI was programmed. They however took no stance wether it was a bug or not.

Kralizec
07-05-2006, 13:44
Fair enough, I retract my statement.

Puzz3D
07-05-2006, 14:49
The load/save was what you say it was, a bug. A bug that took around 6 months of thousands of players playing this game alot to notice.
The AI did have a memory, it just wasn't working properly.
Ever since the bug was discovered I only had 4 seiges broken total, the rest of the seiges during the pre 1.3 patch worked fine.
I doubt CA shipped RTW knowing the bug existed.
Actually, CA said it was a design decision not a bug. After players started complaining about it and their attempt to surpress that complaining failed, they changed it. I've always wondered whether they said it was not a bug to save face. If it was in fact a design decision, then someone at CA knew about it from the beginning.


In a team of programmers, you have to make some ground assumptions so you can get to work. Now that the work is done, you can go back to those hard coded parts of the program and change them. But know full well that for each and every parameter you change, all the other numbers in your program can suddenly start to become buggy, resulting in crashes etc.
If you use top down design, you don't have that problem of changes in one subroutine breaking another subroutine. If you've written spagetti code, then you have a big problem when you make a change. But you missed my point completely. I was responding to your suggestion that modders can fix the game. Well modders can't fix it because of the hardcoded stuff, and CA won't fix it because they are serving the new market not the market that comprises the experienced players. CA has stated that the lack of multiple adjustments on fatigue, ammo and morale is intentional, and the reason given is that they don't want to confuse new players. Apparently, they are obsessed with catering to these new players, and you can see why. New players buy the game. New players become experienced players at some point, but they've already bought the game by that time. So, tough luck if after becoming and experienced player you find that ultimately the game is a shallow experience because it doesn't play well. The problem for CA at that point is to deflect the complaints of the experienced players so that those complaints don't lower future sales.

TB666
07-05-2006, 15:07
Actually, CA said it was a design decision not a bug. After players started complaining about it and their attempt to surpress that complaining failed, they changed it. I've always wondered whether they said it was not a bug to save face. If it was in fact a design decision, then someone at CA knew about it from the beginning.

But wasn't that term a part of the same post where they explained how they AI was programmed ??
I wouldn't be surprised since that post caused alot of noise but as I said before I think it was a simple misunderstanding.
I think they were simply trying to tell us how the AI was suppose to work.
But since the post took no stance wether there was a problem or not it could seem that they tried to brush it to the side.
Also as soon as they moved to SEGA it became a bug and I wonder why *coughs* activision *coughs*.

Puzz3D
07-05-2006, 15:33
But wasn't that term a part of the same post where they explained how they AI was programmed ??
As I recall, they said the AI was reassessing it's options on a reload. I think the problem was that it takes the AI two turns to decide to siege a city. You could find instances of the AI not breaking a siege on a reload, but it was rare.

TB666
07-05-2006, 15:38
As I recall, they said the AI was reassessing it's options on a reload.
Ok then it is the same post I'm thinking of. As I said, think it was a misunderstanding.

I think the problem was that it takes the AI two turns to decide to siege a city. You could find instances of the AI not breaking a siege on a reload, but it was rare.
In my games it was never rare.
In fact the AI breaking a seige after reload was the rare bit here.

A.Saturnus
07-05-2006, 20:26
I think the reason why the save/load-bug was so difficult to confirm, was that the AI behavior on the campaign map seems so random all the time. In all that senseless shoving around of armies the systematicity of breaking up sieges wasn't easy to note.

Laman
07-06-2006, 06:50
One thing to keep in mind is that it is much easier to discover weird stuff in the graphics (though things like the British Legionaries left leg can still slip by) than to see that the ai works well as intended, since you can't test it thouroughly enough, so it can still do idiotic things. And even if you see the ai do some idiotic thing such as suicide charging with their general, it is also easier to see that there is a problem than to actually fix it in a way that works AND doesn't break something else.

Dunhill
07-06-2006, 08:57
But as you note, each of these games only has half the AI requirements of the TW series, which has to have both a strategic and a tactical AI. As for Manassas tactical AI, you don't have multiple maps and terrain to deal with, and I suspect there are just a few unit type, infantry, cavalry, artillery. And BoA doesn't have near as many factions to play (each with their own unique map positions) as the TW series. Nor is there the economic/technlogical development of the TW series in these games.

You're really not comparing the same things here. Each of these games is much simpler than the TW series. Your comparison is simplistic and flawed.

I don't want to come across hard on a old-timer with a join-date as respectable as your's, but I can easily provide an in-depth argument for you.

WARNING: Those of you who are squeamish about in-depth arguments from gorgnards about game's AI's may wish to look away now....

I'm comparing strategic and tactical AIs to the AI used in TW which has strategic and tactical AI. So, my argument is far from flawed, unless you can you show me any intelligent interaction between the AIs in TW? I haven't seen any. One uses the results of the other, but I'd be surprised to know the one provides substantive choice input to the other.

I agree my argument is simplistic as it stands, but that is easily rectified, just a bit boring for those who have played the games I'm comparing. Which you obviously don't considering you only "suspect" TC2M has three unit types. It actually has more, including the officers and supply wagons. In addition, all units can be easily modified, and melee-only units are already being modded. Cavalry can also be mounted or unmounted, and artillery limbered and unlimbered (as well as choosing a number of ammo types). Units can also have a number of formations, including those that result from battle/morale effects.

In each way the games I highlighted are more complex than TW in the areas they deal with. You are correct TW does have a tactical and strategic AI, but I wouldn't say just having both makes the AI better than only having one. What I've said is there are games out there with better AIs, tactical and strategic, because I see no direct connection between the two. I would hope the strategic AI was called during a tactical engagment to ensure quality troops aren't squandered, but I don't see any evidence of that in TW.

The TC2M AI has much more unit information and map information to deal with than TW. The maps include a much greater variety of terrains that effect both movement and fire for units diffenently, and all that can be modded on much, much larger maps. The AI from TW doesnt have much in the way of terrain modification to consider, particulalry when it comes to movement.

The AI from TC2M has to deal with a diverse number of unit attributes that can of course be modded in a large number of ways, including everything TW deals with, but with many more numbers of units on the larger maps, with large numbers of hierarchical leaders, routes for communication, time and weather, LOS, supply, ammo tracking and reinforcements. The player or AI can even request aid in play, which the AI handles, not as scripted events mind you. I don't see the TW AI doing anything like that.

The units positions on the map and relationship within the larger division or corp is also critical to the AI, and how it interacts with units of a different type. Lines are maintined, supply is protected, artillery is set up intelligently. Reserves are kept and used by the AI wisely. Terrain is used to advantage, surprise is possible. Roads are kept open to allow communication between corps, objectives are achieved on map in time and space.

In addition, you can adjust the AI with regard to a handicap for the computer (allow the AI to use bigger and better units), and the amount of CPU processing time you wish to allow the AI to use.

How complicated do you think the tech tree/upgrades/economic model are for the for the strategic AI? I'd suggest they are laughably simple.

If all that happened was for TC2M slapped any old strategic AI on top of thier tactical AI, it would beat TW hands down. As for the numbers of units, as they can be modded at will, the number of unit types is meaningless. TC2M lets you mod at will, including movement (with regard to terrain mods too), does TW let you do that? It all deals with ranged fire, melee, movement, morale, leadership and more.

See how easy it is for me to give a far from simplified, logical, set of statements in support for why the AI for TC2M is much better than the AI for TW? Do I need to do it for COTA and BoA as well?

As an aside, I reckon CA is starting to show up to defend thier product, because they rightly fear the hammering they will get in grognard forums if the AI isn't up to sratch. It's probably not necessary, as someone rightly pointed out the grognards are a small market (although they have some buying power = look at the ASL spending for pristin/unique sets).

sunsmountain
07-06-2006, 11:37
If you use top down design, you don't have that problem of changes in one subroutine breaking another subroutine. If you've written spagetti code, then you have a big problem when you make a change.

No, that is not true. Rome's subroutines depend on numbers declared at the beginning of the program as constants. If you change constants into variables, there are suddenly a lot more configurations of said variables that can have subroutines interfering with each other. It will probably not crash but you have to check before you allow a number to be modded.

All I'm saying is that it takes them time to make a number you want to mod, moddable. Given the fact they were under time pressure, this could mean they haven't made all the numbers moddable yet that they want to make moddable. I do not believe that in principal, they do not want to mod certain numbers, except for: Unit size, number of units on map and in total, number of factions, stuff like that mainly because of resource limitations.


But you missed my point completely. I was responding to your suggestion that modders can fix the game. Well modders can't fix it because of the hardcoded stuff, and CA won't fix it because they are serving the new market not the market that comprises the experienced players.

I understand your point, but disagree: modders can modify overall unit speed, morale and ammo. Fatigue is hard coded, and they can't modify the ratio of cavalry speed to infantry speed, but other than that they can fix a lot. One prime example of this is darth formations, which tries to make use of what is hard coded as good as possible.

What you are implying with the word "fixing" is that they have to change the hard coded number according to your liking, and they're not going to do that. Would you? I think modders have given CA a clear signal on what to make moddable next. In particular unit speed. They might give us some AI numbers to mod. But don't count on it. Rome is very moddable as it. A battlemap editor wouldn't hurt though, in particular one that can set size.


CA has stated that the lack of multiple adjustments on fatigue, ammo and morale is intentional, and the reason given is that they don't want to confuse new players.

And how would you set fatigue then? To Tire Quickly, Tire Normal, Tire Slowly? Or Tire in 3 minutes, Tire in 4 minutes, or Tire in 5 minutes?
20 arrows, 40 or 80?
These options are useless unless you know how long a battle will last on average, so they would have to add an extra Estimated Battle Time.

I can see how that confuses new players.

ghostcamel
07-08-2006, 19:27
Wow.

The mere fact that CA is absolutely aware that about 90% of the fans on these forums would buy MTW 2 if the AI was 10x better and the graphics werent changed at all. 10x improved in everyway, strategic and tatical, i would think that would be impressive. I would rate the graphics of MTW 2 as being about 5x. More than just being 'prettier' or 'higher res' it achieves its effect through increasing random detail, pretty amazing.


I havent really let myself become too excited about MTW 2, it really has ALL the parts to make a classic. Soooo many missed oppurtunities and mistakes and upgrades over the already amazing RTW.... But the glue that holds ALL these parts together is the AI.

As long as CA is aware of how close RTW came to truelly being a classic, in the same way as Shogun was a ground breaking classic in its day. MTW had superior tactical AI but it was really just a refinement of Shogun. Perhaps MTW2, as a refinement of RTW finally will see CA producing another true classic.

I have my fingers crossed.

Servius
07-09-2006, 19:18
One of the initial CA points was that there's no way to showcase the improved AI prior to release. Maybe I don't understand something, but it seems to me that's exactly what a demo would be good for.

Why doesn't CA release a demo with one battle, preset armies, and a map with many different terrain styles on it? It seems to me you'd get a lot of veteran player feedback, for free, rather than having to pay people to playtest the game. Plus, CA would (unless the game's crap) win approval from the vets because they'd feel like (and hopefully have) a real and helpful voice in the fine-tuning of the battle AI.

Again, maybe I've just missed something, but I really don't see the downside of releasing public betas. CA has a large pool of dedicated veteran players here. It just seems silly not to take advantage of that resource prior to a game's release.

Duke John
07-10-2006, 07:37
Because most people who critise the direction of R:TW and M2:TW are not part of CA's targeted customer group? If you release videos with huge bombards, rockets, all kinds of flaming projectiles and fast running units, would you really care what a veteran thinks, or would you be bothered more about wether the video is flashy enough to get the cash from a teenager?

Furious Mental
07-10-2006, 09:44
Well eventually even dumb teenagers have to grow up so the demographic balance should shift in favour of veteran players over time.

Duke John
07-10-2006, 09:51
No, while they will indeed be the next generation of veterans, they will also be replaced by new teenagers with cash to spend. A target group never ages.

AussieGiant
07-10-2006, 11:59
Well one thing you can certainly talk about from the movies are the movement rates. AI will have to wait for the moment...BUT...

if those moments speeds are actually how fast the game plays...

...then your going to need a G suit and a HUD, because it is like watching Apache Attack Helicopters moving around the map with 5 feet of ground clearance.

That will be the beginning of the end as far as I am concerned.

I really hope, in fact I pray that "actual" movement rates will be accurate and that the "Arcade Battle" button will allow the kiddies to play at warp speed.

I mean this is what I would have thought the "Arcade Battle" would be used for so that the actual game is set to realistic setting as a baseline.

Between having a "fast forward button" and the "arcade button", there is more than enough changes available to the "kids" if they want it.

parcelt
07-10-2006, 14:44
I really hope, in fact I pray that "actual" movement rates will be accurate and that the "Arcade Battle" button will allow the kiddies to play at warp speed.

I mean this is what I would have thought the "Arcade Battle" would be used for so that the actual game is set to realistic setting as a baseline.


Now that's a good point. The slightly disturbing thing to me is that its also rather obvious... Am I missing something or wouldn't this be a very simple and feasible solution to a lot of the issues we've been discussing? If CA elaborates the concept a bit, it would allow MTW2 to be shipped almost as two games in one: of course, the fundamentals would be the same (e.g. graphics, AI, etc), but things as running/kill speeds make or break at least the tactical part of the game (I suppose this goes for both veterans as well as the speed-addicts).

If so required (for the marketing people perhaps?), they could even do it the other way around for all I care, and include a 'Realism' button instead of 'Arcade', which would be the baseline then (although I would personally take that as a sad, sad testimony of the downward direction towards Arcadicism the TW series has taken :shame:).

But what I wanted to ask upon reading AussieGiant's post, is whether anybody is aware (I am not) of any statements by CA regarding the use of an Arcade or Realism button in the way described above, i.e. to effectively cater to two very different styles of (tactical) play?

And could this not be further extended (in a sort of options menu?) to include issues on the strategic map such as diplomacy (e.g. complex or not)?

AussieGiant
07-10-2006, 14:57
Hi parcelt,

To answer your question, I am not aware either. I made this statement because I went back and played Rome Total War the other night and saw the "Arcade" button just sitting there.

Then I started thinking about this whole "AI, realism, speed" theme I keep reading here at the .org

Some bright spark at CA had this idea in the first place so it can't too hard to make sure it is actually used to cater for the "two" precieved user groups.

That film clip looked totally "Hollywood" to me and it really has me worried.

As far as I am concerned if CA takes their eye off the "AI, Gameplay" ball too much, they will shoot themselves in the foot.

Graphics are out of date so fast that it would seem like a better investment to be a little off the pace on the graphics side but make sure the game play is still cutting edge.

CA have a wonderful combination of strategy and action in this series but the roots of this game is the "strategy", not the graphics.

Canon's mounted on elephants....I mean really come on:dizzy2:

parcelt
07-10-2006, 15:23
Well, I just took a minute (actually 7, :laugh4: ) to check out the video/trailer.... It starts quite impressive, I like the look of the fortified town.

The battles, however, look like the opposite of what I want to see: knights run into spears, wipe them out in less than a second, repeat, repeat, knights seem unstoppable (and what horses they have, running at the speed of light!) until the other side calls in air support and have their Stealth bombers drop some cluster bombs that wipe out the knights completely (don't tell me it was that big gun, guns that fire something with that effect do not exist today) (that's where clusterbombs come in;).

Even if what some posters say is correct, i.e. that this is just a teaser and therefore not representative of how the actual game will play, the fact that they've chosen to depict a 'battle' at lightning speed says enough about who this game is aimed at. I say this is bad news.

Agree with your comments, AussieGiant. Let's hope CA picked up on this when designing MTW2, and that when shooting that video somebody had accidentally enabled the 'Arcade Extreme' mode....

Puzz3D
07-10-2006, 17:06
But what I wanted to ask upon reading AussieGiant's post, is whether anybody is aware (I am not) of any statements by CA regarding the use of an Arcade or Realism button in the way described above, i.e. to effectively cater to two very different styles of (tactical) play?
Back when the demo of RTW was released, a CA dev answered my criticism of the fast running speeds by saying they had taken a look at the speeds and didn't see anything wrong with them. This was at .com, so the post is long gone. RTW v1.0 had the same movement speeds as the demo. Clearly, the running speeds are unrealistic, but my major concern was that it adversly impacted the gameplay by changing it into a "mastery of the interface" type of game rather than a "mastery of the tactics" type of game. However, eventhough you have more units to control in RTW than in the previous games, CA didn't see anything wrong with the speeds. That indicates they are not concerned that the player has less control.

You can't get anywhere with CA using realism as a reason for a change to the game. Remember, their reason for why the Egyptians are styled 1000 years before the time of RTW is that it's what people expect to see. "People" obviously means their target market. They have made a statement here at the org that gameplay is their primary consideration not realism. However, if they are not concerned about reduced control of the units, I don't know how to take that statement.

What could work to effect change is an appeal to gameplay by trying to show that a suggested change would improve the gameplay for the target market. I can't imagine the target market being adversly affected by fixing the walk/run speed ratios. The problem there seems to be that CA thinks the speeds are ok, and they certainly don't see complaints from the target market about that because the target market doesn't care about it. Just the same, the target market would benefit from the change even if they don't realize it now.

I see no chance of getting two "modes" of play other than what's already there with the on/off switches on fatigue, moral and unlimited ammo. It's simply too much work to balance two modes rather than one. This is also the reason given by CA for there not being more than one era in the game. It's interesting that the earlier games had eras (multiple campaigns), and now we see compromises being made such as "years to turns" to simplify the game into one era. You can't convince me that moving from multiple campaigns to one campaign constitues "improved gameplay" unless you twist the meaning of that phrase.

Puzz3D
07-10-2006, 19:03
In my games it was never rare.
In fact the AI breaking a seige after reload was the rare bit here.
Well it was scientifically studied over at TWC, and shown conclusively to affect the campaign. I even checked it myself with fog-of-war turned off. Virtually every AI siege was broken on a reload. AI factions were hindered in their expansion, and MikeB posted that he tested the campaign with the auto-run feature. He wouldn't see the problem testing it that way.

BTW,there was no misunderstanding. Every thread dealing with this at .com was closed or deleted and people threatend with banning. There was a definite attempt to put a lid on this by squashing criticism and declaring that the game was intended to work this way.

Puzz3D
07-10-2006, 20:34
No, that is not true. Rome's subroutines depend on numbers declared at the beginning of the program as constants. If you change constants into variables, there are suddenly a lot more configurations of said variables that can have subroutines interfering with each other. It will probably not crash but you have to check before you allow a number to be modded.
Fatigue, moral and ammo are not constants. The arcade setting changes them. I'm not talking about making them modable. I'm talking about some ingame options besides on and off for these variables. Ca has steadfastly refused for 5 years to make these ingame options. It has nothing to do with them being constants.



All I'm saying is that it takes them time to make a number you want to mod, moddable. Given the fact they were under time pressure, this could mean they haven't made all the numbers moddable yet that they want to make moddable. I do not believe that in principal, they do not want to mod certain numbers, except for: Unit size, number of units on map and in total, number of factions, stuff like that mainly because of resource limitations.
Well you're wrong. They said they were not going to make these things modable on principle. They said it 5 years ago, and several more times since then. They've had 5 years to provide more settings on the arcade parameters. They haven't, and it's not because they didn't have time.




I understand your point, but disagree: modders can modify overall unit speed, morale and ammo. Fatigue is hard coded, and they can't modify the ratio of cavalry speed to infantry speed, but other than that they can fix a lot. One prime example of this is darth formations, which tries to make use of what is hard coded as good as possible.
Changing overall speed isn't enough. The ratio between walking and running is incorrect. They treated movement speed as an arbitrary variable, and there are plenty of indications that running speeds were increased after the game was balanced, and after the AI was set up for a different speed.


What you are implying with the word "fixing" is that they have to change the hard coded number according to your liking, and they're not going to do that. Would you? I think modders have given CA a clear signal on what to make moddable next. In particular unit speed. They might give us some AI numbers to mod. But don't count on it. Rome is very moddable as it. A battlemap editor wouldn't hurt though, in particular one that can set size.
I'm not implying they make the values to my liking. I'm saying they should put the values back to what they themselves originally determined worked best. You can't change movement speed without messing up the whole system. I think someone at CA doesn't understand this. You can see how they work. They made the maps bigger in MTW, and then didn't adjust the fatigue rate for the larger maps. When we complained that fatigue was too great, we were told is was supposed to be like that so that the battles would seem to be taking place on a larger scale. Later, LongJohn admitted that they had simply forgotten to adjust the fatigue rate for the larger maps. The one concession he made was to reduce the fatigue rate of running cav by 10%.

[QUOTE=sunsmountain]And how would you set fatigue then? To Tire Quickly, Tire Normal, Tire Slowly? Or Tire in 3 minutes, Tire in 4 minutes, or Tire in 5 minutes?
20 arrows, 40 or 80?
These options are useless unless you know how long a battle will last on average, so they would have to add an extra Estimated Battle Time.

I can see how that confuses new players.
The reason we asked for ingame adjustments on these things is because CA was not balancing them. This request occured after STW/MI was released and we worked on the v1.02 beta patch. It was clear that CA was not maintaining the level of balance that they had in original STW. LongJohn had left when MI was balanced. Take a look at the power parameter for guns in MI. Someone increased that from power = 4 to power = 16. They obviously didn't know what they were doing because the parameter saturates at power = 8. Going to 16 doesn't produce any increase in the effectiveness of guns. LongJohn came back for MTW which was definitely better balanced than MI. I don't know who balanced RTW, but they did one lousy job.

You do have to know how long a battle lasts on average to do proper balancing. The purpose of ingame settings would be to allow the players to finish balancing the game since CA isn't doing it. Mods are not the answer for the online community. The options have to be part of the game, and they have to be more than on/off morale, fatigue and ammo adjustments. CA has to fix the running speeds. Every indication is that CA is not doing this, and they already declined to put in the options on fatigue rate, morale and ammo.

BeeSting
07-10-2006, 22:02
No matter how pretty the graphics to enhance the emersion, if the AI still does really stupid things, one will wake to reality and the fun will be ruined. I am all for bettering the graphics, yet I am crossing my fingers that when they say “improved AI” I hope they meant that it has improved as much as the graphic has.

I understand the inability to really show off the AI in their previews, but for the sake of the hardcore fans and just to quench all the negative posts out there that CA is targeting the “dumb/noobish” crowd, I wish for CA to put into words some of the fixes, departures, and improvements they have made from the AI we are all so familiar with in RTW.

When they say improved AI, what do they mean by that? Many are skeptical, yet giving the benefit of doubt, that finally we will have an AI that will at least do some sensible things. For example: will it still make ridiculous and unrealistic diplomatic decisions? Even the imperial Japan thought to offer terms of surrender when their survival was in serious danger. From the countless reviews I’ve read so far it seems the AI will still be suicidal unable to overcome how they feel about a player. Will it still not know that they can't beat a human player with half the amount of soldiers, let a lone with an equal amount? To cater to wide range of player skills out there the AI can remember it's own sucess rates and adjust the size and quality of its force accordingly. Lastly, will it still have suicidal generals?

We are all dying to see the improvement in this game to the level of our satisfaction in their series of games, which we are convinced, is the closest thing out there to making the players feel like he/she is making history. We are concerned however that after all our dying anticipation for their next game, we will still have something that’s put in a nice eye-catching package and yet lacking equal improvement in content, which the game so critically depends on since it is essentially a single player game, the AI. We shall see whether our long patience and loyalty to CA was overlooked for some fast money from new players that may come and go—so signing the end for their niche market in PC games to competing with the many immature market for console based games.

BeeSting
07-10-2006, 22:02
Yikes! Didn't mean to double post

Grifman
07-11-2006, 02:16
WARNING: Those of you who are squeamish about in-depth arguments from gorgnards about game's AI's may wish to look away now....

Indeed. But only because the analysis is so flawed.


I agree my argument is simplistic as it stands, but that is easily rectified, just a bit boring for those who have played the games I'm comparing. Which you obviously don't considering you only "suspect" TC2M has three unit types. It actually has more, including the officers and supply wagons.

So now you have 5. Big deal, TW has generals to match your officers. And as quit obviously, far more units. All sorts of infantry units, some melee, some missile, some both. Same for cavalry. And then there is artillery. And then there are the special units like priests which have minimal combat power but confer other benefits on nearby units.


In addition, all units can be easily modified, and melee-only units are already being modded.

Irrelevant. The AI may not effectively use modded units. That said, the TW engine can be modded to include many more units than in vanilla TW. The point still stands, TW has far more many unit types and capabilities that the AI has to handle.


Cavalry can also be mounted or unmounted, and artillery limbered and unlimbered (as well as choosing a number of ammo types).

The TW engine also included cavalry that could be mounted or unmounted, artillery could be effectively limbered or unlimbered (all this means is that an artillery unit can't move and fire at the same time). And TW has artillery that can move and fire - caroballistas. And TW artillery also has different ammo. Nothing unique here it seems for TC2M at all despite your claim.


Units can also have a number of formations, including those that result from battle/morale effects.

TW units also have different formations as individual units - and also for the entire army. And TW reflects morale also, though it doesn't impact formations that I know of.


In each way the games I highlighted are more complex than TW in the areas they deal with.

So far, I don't see much in your analysis that makes TC2M more complicated. And you seem to have ignored things the TW engine has to do that TC2M doesn't - assault and defend bridges/fords and assault/besiege cities/castles. Hmm, looks like you missed a big gap there, doesn't it?


You are correct TW does have a tactical and strategic AI, but I wouldn't say just having both makes the AI better than only having one.

No, but game development is a matter of resources, and you're being naive if you don't recognize that. Resources - money, time - may not exist to make both best in class.


What I've said is there are games out there with better AIs, tactical and strategic, because I see no direct connection between the two.

Perhaps. But it's very easy to cherry pick what you want. Pick any game and I can just about find some aspect in that game that is handled better in another game. The real issue is does the TW AI provide a reasonable challenge for the majority of players. I'd say that goal is met.


The TC2M AI has much more unit information and map information to deal with than TW. The maps include a much greater variety of terrains that effect both movement and fire for units diffenently, and all that can be modded on much, much larger maps. The AI from TW doesnt have much in the way of terrain modification to consider, particulalry when it comes to movement.

Only if you ignore the need to assault/defend fords/bridges and assault/defend castles/cities. And are you sure about the impact of terrain in TW? Isn't movement different for grass, vs forests, deserts vs hills, vs rivers? Does TC2M have more terrain differences?


The AI from TC2M has to deal with a diverse number of unit attributes that can of course be modded in a large number of ways, including everything TW deals with, but with many more numbers of units on the larger maps, with large numbers of hierarchical leaders, routes for communication, time and weather, LOS, supply, ammo tracking and reinforcements. The player or AI can even request aid in play, which the AI handles, not as scripted events mind you. I don't see the TW AI doing anything like that.

Again, modding is irrelevant since both games can be modded, and we shouldn't be comparing AI with mods, since modding can unbalance a game if the AI can't handle it. That said, I'll grant you some of the above - but you seem to ignore some things in TW.

You ignored the impact of shields/armor and armor piercing weapons. And you've also ignored the entire rock/paper/scissors aspect of TW. Some units defend better against cavalry, others have a strong charge, but have a weaker regular attack, etc. Many units have secondary weapons, some missile, some melee. Different units have different battlefield speed. And TW gives morale based upon location of the commander, location of supporting units, how the battles going, etc.. TW also have more of a paper/scissors thing going on that the AI needs to consider in deciding which units engage/defend against what unit. And again, the TW AI has to assault/defend bridges/fords, castles/cities. TC2M has to do nothing like this that I am aware of.


The units positions on the map and relationship within the larger division or corp is also critical to the AI, and how it interacts with units of a different type. Lines are maintined, supply is protected, artillery is set up intelligently. Reserves are kept and used by the AI wisely. Terrain is used to advantage, surprise is possible. Roads are kept open to allow communication between corps, objectives are achieved on map in time and space.

Yes, here TC2M does have to do some things the TW engine doesn't have to. But it's a different game and the TW AI doesn't have to consider this information. But terrain can be used by TW also - use of hills, forests, rivers, etc.. You seem to be deliberately ignoring things the TW AI has to handle.


In addition, you can adjust the AI with regard to a handicap for the computer (allow the AI to use bigger and better units), and the amount of CPU processing time you wish to allow the AI to use.

And you can't do the same (except for the CPU processing time) for TW? Again, why are you ignoring something the TW AI can do?


How complicated do you think the tech tree/upgrades/economic model are for the for the strategic AI? I'd suggest they are laughably simple.

I'd suggest that's a laughably simplistic view. The strategic AI has to assess risk, invade or be invaded, who do you ally with, who do you declare war on, do you build economic buildings, or religious buildings, or military buildings. Do you build roads or defenses? Do you build land units or naval units? Do you invade by land or by sea? A good strategic AI is very difficult as it has to assess so many possibilities.


If all that happened was for TC2M slapped any old strategic AI on top of thier tactical AI, it would beat TW hands down.

That's a silly statement. A crappy strategic AI would render tactical battles irrelevant. You're hyperbolyzing.


As for the numbers of units, as they can be modded at will, the number of unit types is meaningless.

Not, not really since an AI has to be able to handle them. And we should be talking/comparing vanilla games to be consistent. Mods can change all sorts of things that render comparison difficult.


TC2M lets you mod at will, including movement (with regard to terrain mods too), does TW let you do that?

What are all the TW movement/combat mods all about then? You probably should read up on TW modding before making such statements.


It all deals with ranged fire, melee, movement, morale, leadership and more.

You can mode ranged fire, melee, movement, and morale in TW. Along with armor, armor piercing, missile capabilities and fire speed, capabilities vs specific unit types (cavalry, etc.)


See how easy it is for me to give a far from simplified, logical, set of statements in support for why the AI for TC2M is much better than the AI for TW?

Actually, you haven't proven any such thing. All you've done is given statements showing the complexity of the TC2M AI. You've done nothing (nor have I) to show that the AI is actually any good at considering all those factors and using them effectively in battle. Complexity of AI does not equal better AI.

Duke John
07-11-2006, 07:02
R:TW may have more types of units but it does nothing different between the types. If it's got a ranged attack it will go within missile range and fire away whatever it is, no matter wether it is an onager, archer, javelineer or a caroballista. Before the latest patch or the patch before, missile troops would even charge before the melee units without firing a shot. Infantry units charge enemy units. Cavalry moves to a point to the side of your main body and flanks on its own. The general unit is nothing more than a cavalry unit, how many times have you seen it going after a routing unit to rally, or stand behind the weakest point in the line?

So it's got infantry, cavalry, assorted missile and missile cavalry. 4 types, further differentiated with tagged on extra properties that do not affect AI tactics.

On a higher layer, the AI is unable to form reserves and use them. It does not know how to properly charge the enemy line or even how to keep a steady line when near the enemy. It bunches up its missile units. The AI army is only kept together by strict formations which are dismissed once the enemy is near enough and then it's each unit on it's own.

The bottomline is that the AI does not know how to handle any of its units. CA has admitted themselves that the R:TW AI was a bit of a failure so there is little point in trying to defend it.

Furious Mental
07-11-2006, 08:01
"No, while they will indeed be the next generation of veterans, they will also be replaced by new teenagers with cash to spend. A target group never ages."

You are completely wrong. They may be replaced by new teenagers but over time teenagers will make up an ever smaller portion of the PC game playing community unless people automatically stop playing PC games when they reach a certain age. I do not know what is so hard to understand about this. Take a microcosm of the PC game playing community- ten different people over a period of one century, each one born ten years after the other and starting to play PC games when they are ten and continuing until they die at the age of a hundred. For the first ten years there is only the first player and they are a teenager i.e. 100% of the sample is teenage. During the second decade, the teenager (player 2) is only 50% of the sample. During the third decade, the teenager (player 3) is only 33%. And it continues until by the end of the century the teenager is only 10% of the sample because players 1, 2, 3, 4, 5, 6, 7, 8 and 9 are all aged over 20. Replace "player" with "generation" and that is basically how the demographics of PC gamers will change over time, i.e. with the average age of the PC gamer going up and then plateauing, which will probably be when the first generation of PC gamers are geriatrics.

Duke John
07-11-2006, 08:28
True, but that was not my point. The target group remains. It might get smaller relative to other groups, but that doesn't matter as long as CA thinks that they can get the most sales when targeting young teenagers. R:TW and M2:TW seemed to have been aimed at a younger adience than S:TW and M:TW. The gamer population might be getting more mature on average, but CA's targeting group is getting younger. Instead of waiting for CA to change its marketing strategy I would much rather see other companies filling the gap.

Furious Mental
07-11-2006, 09:59
Yes well I think that will happen if CA decides they will keep targeting a demographic that gets proportionally smaller over time while the number of grouches who want artificial intelligence rather artificial mentally defective intelligence grows proportionally larger.

AussieGiant
07-11-2006, 12:05
I would have thought that game speeds getting faster would be counteractive to the teenagers.

It certainly makes the game harder, and if a teenager loses more than twice you have lost their interest. Of course the arguement (a bad one at that) is that if you get to this stage CA has already got the kiddies cash so they don't care.

Of coruse again, that is a short term view, as the kiddies will remember for next time and say it was too hard and not buy the next version.

I wold love to know what CA's thought process is about NOT keeping movement rates accurate?

Does anyone know?

Duke John
07-11-2006, 12:32
It looks more exciting. If they used accurate movement speeds than the units would seem to run slow, especially when you are used to the speed of units in standard RTS games. However if you used the speed of R:TW units in RTS games then it would look hyper again.

Thing is that most RTS games have units composed of just 1 soldier or a small group of soldiers. To have that unit run 4 times its size (draw a circle around all the soldiers, that is the size) could take 5 seconds. To have an unit in TW, which contain many more soldiers, run 4 times its size might take 25 seconds. To some people that might appear slow, while in fact the soldiers are running at the same speed as in RTS games. The fix was apparently to increase the running speed of the soldiers in R:TW.

Compare it with dropping a small wooden box from 10 metres. It appears to be falling quite fast. Then imagine dropping a hangar from 10 km. And view it from a large enough distance to make it appear having the same distance as the wooden box. Now the hangar doesn't even appear to be falling, while it is still falling at the same speed and acceleration as the wooden box (ignoring wind drag).

This exageration of physics is quite common in games and movies alike. Just think of how people get blown back by a shotgun or even a regular bullet.

Well, that is just my theory :wink:

Dunhill
07-11-2006, 12:40
Crikey mate, good try but you are out of your league (like Australia in the World Cup)

I put forward a number of valid points to suggest that the AI could easily be better, and support this by citing examples from a recent releases by INDY developers. It is a strong argument, with few flaws.

The TW AI doesn't process on-field supply, or the effect of multiple leaders with a variety of leadership characteristics that effect gameplay, and with different spheres of command. I forgot to include the couriers which actually fulfill a function on the map and are run by the AI ( to account for terrain and prior courier events and enemy units)

To say the AI can't make use of the information ascribed to a unit is rather poor for the AI. I'd have hoped the AI knew full well the general attributes of the units it had at its disposal. I'd have thought that would be at the core of the AI, knowing what a unit can do and how to use it?

""Hmm, looks like you missed a big gap there, doesn't it?""

The gap was never there. It was my fault for not pointing this out. TC2M does have fords, bridges, fences and other obstacles to deal with, as well as the movement of many more units over much larger maps. Leaders and thier interactions with couriers and other leaders and units like ammo wagons add complexity to the TC2M battle that I don't see in TW. The AI puts units in reserve, calls them up when needed, moves wagons to supply ammo, and pulls units out of the line as required to conserve thier effectivenss and allow them to get more ammo. A short battle in TC2M takes around 45 minutes and that's usually just a division with 4-5 brigades of infantry, a couple of units of artillery and cavarly, plus ammo wagons and all the leaders. Once things move up to the corp and army levels, things get very complex and battles can last 3-4 hours. easily.

TC2M's AI makes use of favourable terrain for movement and combat, as well as surprise (it can handle siege across rivers). The existence of bridges, fords, breastworks, and various types of housing and forest/fields/rough enable the AI to actually make choices in the game about how to move large groups of brigades/divisions/corps. The terrain effects are more pronounced in TC2M. Particulalry with fire, movement and fatigue. There are 16 terrain factors, plus weather and light effects for the AI to compute in the vanilla version. It may be comparing apples and oranges considering the size of the TC2M playing field and the ability to restrict views to what your units can see. The AI may have more scope for using terrain for a surprise that TW doesnt have.

There are only two main developers working part-time on TC2M, and they've produced a game with better tactical gameplay than the current TW tactical AI. I'm not the only person saying this. Many of the people who have tried the game indicate that it is more immersive and has a very good AI. Its an INDY development, built by grognards for grognards. I must admit they spent most of thier development power on the AI, so it may be unfair to compare it to TW, which tries to do lots of other things.

""The real issue is does the TW AI provide a reasonable challenge for the majority of players. I'd say that goal is met.""

I'm comparing what the AI's have to handle to do the job. Sadly, the TW AI doesnt provide a reasonable challenge for many of us (but not most of us), and hasn't done for some time. If the goal is to please the most people, then I agree the TW AI is fine, but its not for the veteran player anymore. On that point the goal has not been met for some time.

I'm saying that TC2M is open to more modding than TW, but the AI isn't open to modding. So it is essentaily irrelevant to this discussion of the AI.

The AI for TC2M is dealing with so many more variables and doing it in a way that is enables it to beat veteran TW players. I'm not the only TW veteran to be praising this AI. Other INDY game developers with very good AIs are praising thier efforts with this AI.

I am not ignoring those things that TW does, just showing the number of things that the AI for TC2M is doing that contribute to a better gaming experience for me. The difficulty setting for TW doenst do much to make for a better game.

I see the strategic side of risk analysis as a quantitative operation, which isn't as difficult to model as the tactical situation which changes significantly in real time over space. These attributes make it much more difficult for the AI. The turn based nature of the strategic AI enable the risk analysis to proceed without real time changes in space. Part of this analysis can also be qualitative. Crikeys mate, you've got me arguing about the differences between qualitative and quanititative risk analysis after work. I'm going to have to start charging you my professional training/consulting fee.

The lack of a strategic AI for TC2M is almost always brought up by game reviewers. I don't see it as exageration to say that adding a, shall we say "standard", strategic AI would put this game in the same league as TW. I'm sure the reviewers would make an even more direct comparison.

I don't see any change to the battle speed that the TW veterans have been complaining about and requesting repairs for. I suggest you have a look at some of the other threads about battle speed.

""You've done nothing (nor have I) to show that the AI is actually any good at considering all those factors and using them effectively in battle. Complexity of AI does not equal better AI.""

You are correct, I've given quite a number of reasons why the AI for TC2M is more complex. I can only base my rating of the AI on its ability to give me a good game. The TW AIs don't give me a good game, and haven't done so since STW. Judging from the significant number of posts I'm not the only one concerned about the AI. Conversely, the AI of TC2M is the subject of much praise, particulalry amongst those people who've played TW. I can only judge from my enjoyment of the game, it's surprising and enjoyable to play against the AI from TC2M. I can't say the same for RTW. I just think it is the complexity that is allowing the TC2M AI to beat me. It may be the ability to devote more processing power to the AI.

I think the AI for TW could be much better, and that there are games available that support my argument. I rest my case.

PS: I write complex legal and scientific argument regarding regulatory decisions that I make on behalf of my government on veterinary medicines all day. I specialise in qualitative and quantitative risk analysis of toxiciology, trade and health impacts of the new chemicals I'm asked to evaluate.

However, this doesnt mean I enjoy spending my free time writing arguments (as indicated by my relatively low post count here). I'm an Australian and I''d rather be watching the footy instead of doing something that feels more and more like what I do at work. Blah, blah, blah, blah, blah, good God I've even bored myself silly.

https://forums.totalwar.org/vb/images/smilies/gc/gc-wall.gif
:wall:

Take it easy mate. You'll have to buy me a slab of Boags Premium and come over to listen to the footy some afternoon if you want me to continue yammering on about this subject.

Puzz3D
07-11-2006, 12:54
I would have thought that game speeds getting faster would be counteractive to the teenagers.

It certainly makes the game harder, and if a teenager loses more than twice you have lost their interest.
In RTW/BI in the tactical battles, the AI will charge with units are weaker than yours. This makes it easier for the player when he has the better units than it was in the previous engine where the AI would not make frontal assaults with weaker units. On the strategic map, the AI will attack your army when its army is weaker. So, as long as you move up the tech tree faster than the AI, you'll be ok in the battles since they are mostly determined by frontal fighting. Battle speed can be fast if they are just a matter of slamming the armies together frontally. It's moving up the tech tree to get the better units that's important. The difficulty settings are there for players who find medium difficulty too easy. They give a combat boost to AI units and reduce the money a player has to start.

It seems to me that the emphasis in the series has shifted away from using maneuver to win battles towards moving up the tech tree as fast as possible. Since the battles are actually too easy, the game now has a bias favoring the AI in the auto-resolve to force the player to play out the battles.

O'ETAIPOS
07-11-2006, 13:45
It seems strange to me how some people tend to defend RTW AI - "I know it has some flaws but look - it is so massive it has bidges, sieges, camp and tactic AI" this is NOT an argument.

No matter how big it is, if it's not working correctly it's useless. Think of a ship with small flaw - one plate on hull is missing. Is it a good ship? Even if it is very big one?

Some say time and resources are limited. They are. When you build house and you know you have limited resources do you start building front wall up to 5th floor leaving back at ground level? I think nobody would buy house with beautiful front and nothing behind.

And this is what we were offered by CA in RTW - Beautiful front up to 5th floor (graphics) and ill built, unfinished up to 3rd floor in the back (AI).

sunsmountain
07-11-2006, 14:14
@Dunhill

I see the strategic side of risk analysis as a quantitative operation, which isn't as difficult to model as the tactical situation which changes significantly in real time over space. These attributes make it much more difficult for the AI.
This is reason enough for Take Command: 2nd Manassas not to be compared to Rome TW. Why you're bringing this TC2M into the discussion, when most of us do not have a clue what you're talking about, just to make your point (which seems to be that its possible to create a better AI - for which TC2M is not an example), is a little beyond me. Going on about the TC2M engine and arguing against us doesn't hide the fact that you know quite little about the Rome:TW engine and have little to add about it, despite the length of your posts.


The turn based nature of the strategic AI enable the risk analysis to proceed without real time changes in space. Part of this analysis can also be qualitative.
For the Rome engine, getting any risk analysis at all is difficult, let alone what to do with it.


PS: I write complex legal and scientific argument regarding regulatory decisions that I make on behalf of my government on veterinary medicines all day. I specialise in qualitative and quantitative risk analysis of toxiciology, trade and health impacts of the new chemicals I'm asked to evaluate.
In general, and certainly not on a gamers forum, do people like to hear about how great you are. I certainly do not. It does explain the length of your posts, and your desire to get some recognition here.

Granted, my posts can be long as well, but that's usually because i'm responding to long posts or actually have something to say about the engine of a relevant game to MTW2, ie RTW or STW/MTW.


Indeed. But only because the analysis is so flawed.
Tell 'm, Grifman. I've about as much patience with TC2M fans as I have with LOTR:BFME fans on this forum: NONE.

sunsmountain
07-11-2006, 14:24
@Puzz3D:


Fatigue, moral and ammo are not constants. The arcade setting changes them. I'm not talking about making them modable. I'm talking about some ingame options besides on and off for these variables.
That's kind of contradicting, on the one side you want more than on/off, but you don't want them moddable. Sounds to me like you do :wink3: . When i mentioned constants, i wasn't referring to fatigue, morale and ammo. I was referring to the change in status a moddable variable gets with regards to an non-moddable one. These moddable variables have to be tracked and checked. You can't just free them up, it screws up the whole battle engine if people don't know what they're doing while you are allowing them to change it. I think that is why CA won't let people change it.

Should they be moddable? Hell yeah.
Should they be moddable through an ingame menu? Hell no, the balance should be right to begin with.


They said they were not going to make these things modable on principle.

Changing overall speed isn't enough. The ratio between walking and running is incorrect. They treated movement speed as an arbitrary variable, and there are plenty of indications that running speeds were increased after the game was balanced, and after the AI was set up for a different speed.
For both statements, I'm asking you for something of a link or a reference please.

On the second, if movement speed is an arbitrary variable that can be changed, it can also be changed back to its original value. The AI you're referring to seems to be set up for the old STW/MTW speeds, and while I would like to believe they copy-pasted some code of that old engine, I know (and have quotes to prove) that they didn't and don't.


I'm not implying they make the values to my liking. I'm saying they should put the values back to what they themselves originally determined worked best. You can't change movement speed without messing up the whole system. I think someone at CA doesn't understand this.
I agree but they did those things for a reason, not out of stupidity. And you know the reason: Motion captured animations. They have to be in sync. That is the starting point. The new battle engine will have to be built around that, and not about STW/MTW battle speeds or AI.


You do have to know how long a battle lasts on average to do proper balancing. The purpose of ingame settings would be to allow the players to finish balancing the game since CA isn't doing it.
I doubt the average player can balance the game better than CA can. I sincerely hope CA will take into acount the average battle time length, and the dependent variables battle map size, unit size, morale + expertise of general, ammo, fatigue + effect of weather, infantry/cav movement speeds, etc.

and make this moddable. So we can simply choose whether we would like fast or long battles. AI decision making should be good regardless.

Puzz3D
07-11-2006, 16:35
That's kind of contradicting, on the one side you want more than on/off, but you don't want them moddable. Sounds to me like you do.
I just explained to you why ingame options are not the same as making something modable.



When i mentioned constants, i wasn't referring to fatigue, morale and ammo. I was referring to the change in status a moddable variable gets with regards to an non-moddable one. These moddable variables have to be tracked and checked. You can't just free them up, it screws up the whole battle engine if people don't know what they're doing while you are allowing them to change it. I think that is why CA won't let people change it.
Don't make me laugh. They don't track and check modable variables. All they did was free them up. You can plug in values to those modable variables that are out of range.



Should they be moddable? Hell yeah.
Should they be moddable through an ingame menu? Hell no, the balance should be right to begin with.
Yes it should be right to begin with. Fat chance of that happening.



For both statements, I'm asking you for something of a link or a reference please.
They have made ammo and morale modable, but not fatigue rate although you can affect this with teh stamina setting. I actually meant provide ingame options on those settings which they decline to do on the principle that it would confuse players. I don't have a link for that. It's been going back and forth for 5 years in writing, verbally and possibly once in the forum. So, they agree with you that ingame options are bad for sales.

For the second, I posted the reasons in this forum why I think running speed was changed late in the development, but it's within some other thread.


On the second, if movement speed is an arbitrary variable that can be changed, it can also be changed back to its original value.
The key word is arbitrary not variable. When they treat movement speed as arbitrary in a game that's supposed to be to scale it shows that the game is no longer based in reality. They have stepped over the threshold into blatant fantasy that screams at you every time you see a unit run. There is now reason to believe that M2TW will reflect an even more distorted version of reality than the previous games. Total War has become The DaVinci Code of historical gaming.


The AI you're referring to seems to be set up for the old STW/MTW speeds, and while I would like to believe they copy-pasted some code of that old engine, I know (and have quotes to prove) that they didn't and don't.
The new engine doesn't allow you to change walk, run and charge speeds independently. That was the very first thing I tried to do when I got the game. I certainly had enough experience with the system to immediatly understand there was a problem with movement. The running speeds are ludicrous looking, but the real problem is how they detract from the gameplay. The battles are now a game of "master the interface" instead of tactical sparing where you have time to respond to various threats and the interface plays a minimal role.

They actually used too much of the old AI in RTW. The old AI isn't suited to using a phalanx, and the RTW AI can't either. One of the chief criticisms of RTW is how the AI doesn't maintain a battleline. It's still using the old AI technique of making individual unit matchups without regard for the unit's flank exposure. This is catastophic for a slow phalanx which has very little power except to the front. I've played mods of RTW, and none of them are able to fix this inept use of phalanx by the AI.



I agree but they did those things for a reason, not out of stupidity. And you know the reason: Motion captured animations. They have to be in sync. That is the starting point. The new battle engine will have to be built around that, and not about STW/MTW battle speeds or AI.
Motion capture is fine, but they didn't scale it to a reasonable speed for the type of battlefield units involved. Practically all the RTW/BI mods slowdown the units. Either whoever chose the speed didn't understand the impact it would have on the gameplay or they did understand and chose the speed for that reason. Of course, CA isn't going to say they didn't understand even if that's the case.



I doubt the average player can balance the game better than CA can. I sincerely hope CA will take into acount the average battle time length, and the dependent variables battle map size, unit size, morale + expertise of general, ammo, fatigue + effect of weather, infantry/cav movement speeds, etc.

and make this moddable. So we can simply choose whether we would like fast or long battles. AI decision making should be good regardless.
Great, but that doesn't help me going online. If there were some options on these things besides on/off, someone could host a game with settings that does improve the gameplay over that which CA deems acceptable. It's relatively easy to improve the gameplay over the official gameplay, although, it's hard to bring the game close to its full potential since that requires very good balancing of many parameters. Unfortunately, balance also includes balancing the unit types, and CA has chosen to include so many units types that they can't possibly balance them in the available time. In addition, there are the missing features, which detracts from the gameplay, in the new battle engine that were in the old battle engine.

Dunhill
07-12-2006, 09:06
I'm going to extricate myself from this mire of this debate. It's not worth my time.

Puzz3D does a better job than me of handling the long posts. I reckon CA wouldn't go far wrong hiring him as an adviser. He's got the subject covered.



I'll be watching the The Wallabies or the All Blacks mash up some poor buggers.

Crian
07-12-2006, 13:45
Ok, I'm no programmer alright? But I think Dunhill has made a very good argument that the TW AI has so much more potential and the devs at CA aren't doing enough to make it better.

A lot of people here have been defending (wait, you're not employees right? kidding :laugh4:) CA by saying "Total War is so complex", "There are a thousand things for the AI to consider", or by saying things to people like "You don't know how the engine works" or "You're not a programmer", "CA knows better", etc. In my opinion, these are all just lame excuses. People, the Total War AI sucks, we all know it, and (again, I'm not a programmer) while I think we all agree that they can't produce an AI that can perform like Hannibal or Alexander, there has to be SOMETHING they can do.

For instance, on the Strategic part of the AI, how hard could it be to tell the AI to move its armies together and not in tiny bits? Or treat the human player with more consideration (ie. Be more cautious; when at war with the player focus all its resources against the player, etc.)

Or better still, why can't they script the AI to behave a certain, predictable way just to "appear" that it's thinking? I remember how the Danes in MTW remain independent despite the fact that the HRE can take it over instantly, or how the Sicilians manage to launch dozens of ships EVEN if they only have one stinking province, or how the Papacy remains independent (most of the time). What I mean is, there has got to be some type of scripting behind that right? Overall, it gave the factions a lot more "flavor", and the appearance that the AI thinks, that that AI factions have some sort of relationship. A lot will say that the Rome and Medieval 2 campaign maps are much more complicated than the old MTW map, but seriously, do you believe that they can't pull that off? I don't use diplomats in RTW, heck, I don't even move my first diplomat regardless of what faction I play. Why? Because when you go right down to it, it's useless, and you can win the game without allying with anyone and without bothering to talk with anyone. You all KNOW what I'm saying is true. That's how useless the game's AI is, so how could anyone here defend it?

In the battlefield, how hard could it be to prevent the AI from doing stupid things? Many keep arguing, oh there's a hundred unit types, there's terrain, weapon types, armor, fatigue, weather, etc... You know, a lot of this is not really as important as some might consider.

When in the battlefield, not all of this needs to be taken into account ALL the time. There are just a few basic unit types, and that's all the AI really needs to take into account. Like for example, Roman Archers and Archer Warband. How different are these two units? Take Legionnary Cavalry for example, and Cataphracts, how different are they? Sure there are a lot of details that set them apart but take a good look people, they're BOTH heavy cavalry. What I mean is that every unit can be categorized under a certain type, now this doesn't mean that it is exactly the same as another unit in the same category, I don't mean that, BUT what I mean is that it has essentially the same function as that unit, regardless whether it will perform better or worse. Like, how would you use Poeni Infantry and Hoplites for example? Your main battle line, right? Because they essentially have the SAME FUNCTION regardless of morale, fatigue, armor, weapon ... whatever you will STILL use it the same way. If you have, say, gold chevron archers with full armor and weapon upgrades, will you use them as frontline shock troops against say a rookie warband with no upgrades of any kind? Still no (Ok I don't know about you guys but I know I won't, it just doesn't feel right :laugh4:). Here's another battle scenario, If you have a rookie unit of Hastati and a gold chevron general unit with upgrades, for example, and you are faced with a rookie phalanx unit, which unit will you send to hold the phalanx down? The general? No, the Hastati, and you will send the general to the flank or rear. You could argue that I'm generalizing and you're right... But it's the AI I'm talking about.

What I mean is that every unit can be assigned a role, and be told specifically what it should do, most of the time. Like don't send your archers into melee, and keep you main line together, always. Don't charge cavalry into spears, etc. These things can be quantified, right? Like when the value of a unit is 1, with 1 meaning it's a missile unit, then don't make it charge unless the battle is being lost. Yes it's simplistic, but I really believe it can be done. Like I said earlier, I'm not asking for a brilliant, dynamic AI, but at least one that doesn't do stupid things and at least APPEARS to be using its units right.

Puzz3D
07-12-2006, 17:16
Ok, I'm no programmer alright? But I think Dunhill has made a very good argument that the TW AI has so much more potential and the devs at CA aren't doing enough to make it better.
The Total War battlefield AI does a couple of things very well. It chooses good melee matchups, and good targets for its ranged units. It always did that even back in STW, and since STW had the strongest rock, paper, scissors (RPS) of any Total War game and expensive units like warrior monks had low armor, the AI gave you a good fight. It's my impression that the STW AI was better than the RTW/BI AI at not firing into a melee when friendly units were present. That's not to say the AI did well in skirmishing its ranged units against enemy ranged units, because it didn't. The AI moved it's ranged units around too much which prevented them from shooting. It still seems to have that problem to some degree even in the latest version RTW/BI, and a new problem arose with ranged units entering into melee before using all of their ammo. I believe this is a consequence of lowering the melee combat threshold which the AI uses to trigger a unit to attack. This is probably also why generals suicide against units they can't beat. The threshold should be raised back to where an AI unit doesn't attack if it doesn't have an advantage.

Since STW, some features were added to the units, but the AI hasn't been made aware of them. For instance, some units have shields that protect against projectiles, but the AI doesn't seem to understand that and, after advancing, a unit will often change its mind, turn around and walk back to its lines allowing the unit to be decimated by projectiles from behind when the shield was giving adequate protection as long as the unit faced the enemy. While the AI does make an estimate of its chances in the melee, it doesn't seem to make any assessment of its potential losses due to projectiles. The Total War battlefield AI has never done this as far as I can tell which is probably why it has always moved its range units around too much. It doesn't realize that enemy ranged units can cause it casualties. You could see this clearly in RTW when AI units would stand outside a city and loose every man to projectiles from the walls.

Phalanx has been added which is very vulnerable to flank attack, but the AI uses the phalanx individually as it would non-phalanx units. That's a unit that needs an enhanced AI so that it will maintain a battleline with other phalanx units or have lighter units protecting its flanks when engaged. Shield wall is a similar feature which is most effective if a battleline is maintained.

For the Total War AI to improve, it needs another level of complexity where it continuously evaluates the position of each unit relative to the other units in the whole army formation. This would give it the intelligence necessary to maintain a battleline and utilize a tactical reserve, and also allow the AI to integrate reenforcements into an army that's already on the battlefield or coordinate two armies.

The AI also needs to be made aware of fatigue and how to manage it. Large maps increase the importance of managing fatigue.

Doug-Thompson
07-12-2006, 17:41
Surely the difficulty of crafting a killer AI increases greatly every time some good new feature is added, such as better diplomacy or tactical options for the battelfield AI.

My apologies if that point has already been made. I've just come back to the forum after one of my long absences.

Also, I think graphics are an important selling point to potentitial new customers.

People with a high, historical interest in the era or who love strategy will buy this game. Really, there aren't many strategy options unless you want to perform magic, fly through the galaxies or some similar nonsense. Marketing is right to market to target potential new customers.

Spino
07-12-2006, 22:53
The need for CA to craft a 'killer AI' has never been a primary concern of veteran TW gamers; most of us would have been quite content with a 'competent AI', one that simply gets the basics right and avoids making ridiculously stupid decisions. Regarding AI Medieval seems to be the only TW title that provides the most consistent challenge on both the tactical & strategic level (especially after modders figured out how to change the AI's build priorities for troops & buildings). Many of us were a tad disappointed that CA & Activision failed to make the improvement of Medieval's AI a priority in the patches or the expansion pack. We didn't think the AI needed a massive boost, only a bit of tweaking to make it 'good enough'. Rome's AI was a full blown step backwards which was all the more shocking considering that the tactical battles (as related solely to unit data & types) were basically the same beast as in Medieval.

People here keep touting the strengths of the AI in the Take Command series, specifically TC2M. It's true that the RTW tactical AI needs alot of work but it's strategic AI needs a complete overhaul; it's simply awful. Understrength stacks led by family members, full stacks led by nameless captains, it's penchance for engaging in multiple front wars, etc. Given that MTW2's strategic gameplay is looking to be even more complicated than Rome's CA had better move heaven and earth to bring the strategic AI up to snuff.

If you want a perfect example of a strategic AI done right check out Galactic Civilizations II by Stardock. GC2 has the best AI of any Civilization style strategy game I've ever played and it doesn't cheat unless you jack up the difficulty levels to where it receives handicap bonuses. I'll go one step further and say that GalCiv2 has one of the best AI opponents ever to be incorporated into a computer game. And yet even after two major patches Stardock is still tweaking the AI so as to improve gameplay and keep veteran gamers coming back for more. Why is it that small, independent developers like Stardock & Madminute Games can create competent (if not downright clever) AI opponents? Perhaps these companies keep a few Nobel Prize caliber programmers on their respective payrolls they're not telling anyone about. Possibly, but I believe it's because these developers made the creation of a strong AI opponent a priority.

Here's a post from one of GalCiv2's developers from the official forums. The subject heading tells you all you need to know about where this developer's priorities are...

http://forums.galciv2.com/index.aspx?ForumID=164&AID=122241


Why AI must evolve after release

To keep the game fresh 6/30/2006 7:35:48 PM

So you've bought the game and you really like it but then time passes and it starts to get old. It's not that the game is boring or bad but that you have figured out how to "beat it". You crank up the difficulty level but then it's ridiculously hard to the point you know you have no chance.

If the game has multiplayer, you play people on-line but as you get further away from the original release, the harder it gets to find people and in particular the harder it gets to find normal people -- people who play the game as it was designed to be played as opposed to exploiting every minor game mechanic flaw to their advantage which tends to suck the fun out.

So ends the typical strategy game. It's usually a question of when, not if, the game reaches the end of its playable lifetime.

When we made Galactic Civilizations II, we budgeted some money for after-release work. We did go a little overboard with that with 1.1 and 1.2 in that the guys just never came down out of "crunch mode" and continued to put in crazy amounts of hours to deliver immense numbers of changes. Now things have returned more to a normal pace and so we look forward to 1.3.

Version 1.3 of GalCiv II will largely involve computer AI work by me. It's not just about making it tougher as much as it is about making it play differently and more intelligently. It will also likely involve creating additional difficulty levels and fixing the $@#$@ thing that causes some AI players to be set to stupid while others are at hard in order to average things out (i.e. someone setting the overall difficulty level to "Painful" only to find that 3 of the AIs are set to being really hard and the other 5 are set to being brain dead -- your best solution for now is to set the players you want to be intelligent or better if you're looking for a challenge and control the players rather than having the overall difficulty system try to guess what kind of gmae you want).

I have a lot of ideas on how to make the game play better, particularly with regard to the CPU option for the AI being able to use better algorithms. The expansion pack, Dark Avatar, will have a LOT LOT more of this kind of thing. But my job is to make sure that the AI continues to provide an interestig experience for players during the summer months.
:jawdrop: ~:shock: :2thumbsup:

GalCiv2 is already a fantastic game and is doing remarkably well in sales too boot and yet Stardock continues to patch the game and add tons of free content and improvements with each patch. What's even more astounding is that by 'industry standards' the game is already good enough (and certainly challenging enough) for the casual strategy gamer but the developer is adamant in its commitment to continue to fix and improve the game so as to satisfy even the most skillful of players. Talk about a shining example of game development & publishing.

Doug-Thompson
07-13-2006, 03:19
I've not played Galactic Civilization II, but do remember the Gamespot review absolutely raving with praise about the AI when it first came out. Called it the best available, IIRC. OK. I'm officially jealous.:embarassed:

However, I will point out that the "stupid" AI does mean more battles.

That's not a compliment. I'd rather have one good fight against a tactically competent AI than several hundred hard but stupid horde-killings at a river crossing.

However, I think the AI of RTW attacks everybody, whether it can afford to or not, for a reason. The same with attacking in penny packets of bad units.

A conscious decision was made to greatly increase the pace of the game from MTW1. We all noticed the ridiculously fast-paced tactical battles. The strategy game suffers from the same flaw. More precisely, it suffers its own flaws that stem directly from the core decision to make the game "more exciting" at the expense of logic.

I'd argue that this was a very bad decision, but that's different from calling CA designers idiots. I don't think they're insensitive to consumer demand. I think they tried to attract new customers — admittedly, at the expense of their steady ones — and have paid a big price when the fan base revolted. I think that's part of the reason we're seeing the Medieval period again.

Puzz3D
07-13-2006, 03:58
People with a high, historical interest in the era or who love strategy will buy this game. Really, there aren't many strategy options unless you want to perform magic, fly through the galaxies or some similar nonsense. Marketing is right to market to target potential new customers.
I think they are in danger of loosing the people with a high historical interest and a love of strategy games. These are the people from which the complaints are coming.

Crian
07-13-2006, 06:18
It would be wrong for us to compare independent developers (ie. Stardock and MadMinute) to CA. :no:

I've played Gal Civ, and yes it is fantastic. I've played both GalCiv 1 and 2, and the AI is really incredible. BUT the truth of the matter is that the learning curve is pretty steep, even when you tone down the difficulty. Games like these target a niche audience, and they satisfy that niche audience by delivering quality gameplay. Total War, on the other hand, is going mainstream, targetting the much, much bigger consumer market. When I first played GalCiv1, it took me more than a dozen restarts to win my first game, because I was used to games being so easy to beat, and this is at the easiest difficulty! After I got that the hang of it I started winning more often... but my point is, had I been a "mainstream" gamer, either I 1.) would not have even given GalCiv a look because of the graphics; or 2.) didn't last 10 minutes playing it because I simply had no patience. I was a bit frustrated at the difficulty at first, but I wanted to keep playing, I was enjoying it! A lot of gamers don't have an imagination to compensate for the lack of visuals nor the patience to learn the nuances of a deep game, that is the sad fact, and CA knows that. The majority rules. :wall:

I have a friend, an FPS and RPG fanatic who has no patience at all with strategy games. He saw me playing Shogun and I tried to impress him by showing him how much you can do, that you are like a real warlord managing your own small empire, how much tactics are needed to win a battle etc.... BUT he can't get over the graphics!! :furious3: He kept saying "What kind of game is that? A contest on who has the prettier formations? Those soldiers look awful!" I wanted to bang his head on the monitor. Now... 3 years later, same person sees Rome, and he goes "OMG! You gotta let me play that!!!" :dizzy2: Sure, he's just one guy, but then I can name several other friends who had almost the same reaction, and I'm sure most of us can relate...

Now... a different friend of mine (also an RPG and FPS person) started playing Rome a few months ago. He was impressed with the graphics as usual and was having a blast playing at Medium/Medium (at this time I have already downloaded mods for Rome and STILL can't enjoy it). After about a week, he finished his first campaign, and then he was back to playing his shooters. I urged him to try another faction or crank up the difficulty, but he didn't want to play it again. Why? He said it was TOO EASY and the end of the campaign sucked (I agree with him on both accounts), and wouldn't want to bother with it anymore. Now... I wonder how many non-veterans had the same reaction to Rome after one campaign? I'm sure there are quite a handful.

CA, you can have your fancy graphics, just make a GAME... not an interactive 3D animation... 'cause at the end of the day, whatever you choose, you will still reap huge amounts of money... but end up satisfying no one if you don't heed calls for better gameplay. The shooter/clickfest market you're trying to tap will buy your product, take a look at the pretty scenery, then lose interest. Or... they could become one of us, loyal fans and veterans who feel that they've been had. :no: :shame:

parcelt
07-13-2006, 09:28
Wow, the latest few posts contain some really interesting thoughts/good comments:

(Gal Civ programmer quoted by Spino)

I have a lot of ideas on how to make the game play better, particularly with regard to the CPU option for the AI being able to use better algorithms. The expansion pack, Dark Avatar, will have a LOT LOT more of this kind of thing. But my job is to make sure that the AI continues to provide an interestig experience for players during the summer months.
I didn't know this, it almost brought tears to my eyes to see such a statement by a programmer! So it ís possible, they dó exist!! (games that are being cared for, with AI as top priority). I think there's a lesson to be learned here by a certain somebody (game company)...



CA, you can have your fancy graphics, just make a GAME... not an interactive 3D animation... 'cause at the end of the day, whatever you choose, you will still reap huge amounts of money... but end up satisfying no one if you don't heed calls for better gameplay. The shooter/clickfest market you're trying to tap will buy your product, take a look at the pretty scenery, then lose interest. Or... they could become one of us, loyal fans and veterans who feel that they've been had.
This is very, very true IMHO, and this is the message Sega's marketing folks need to pick up on.



However, I think the AI of RTW attacks everybody, whether it can afford to or not, for a reason. The same with attacking in penny packets of bad units.

A conscious decision was made to greatly increase the pace of the game
from MTW1. We all noticed the ridiculously fast-paced tactical battles. The strategy game suffers from the same flaw. More precisely, it suffers its own flaws that stem directly from the core decision to make the game "more exciting" at the expense of logic.
I hadn't considered this before, it's an interesting line of thougth.... Though I am not sure yet what would be more disturbing: bad strategic AI being the cause of the mini-army flea attacks, or that being the result of a concious design decision? Either way, the result has been a terrible and saddening weakening of the game's strategic gameplay.


And to Crian, I could quote you're entire post as it makes perfect sense to me and I fully agree. Not being a progammer either (anything but), I still refuse to believe that the type of common-sense improvements you're suggesting could not be (quite easily even) implemented.

econ21
07-13-2006, 14:56
Regarding AI Medieval seems to be the only TW title that provides the most consistent challenge on both the tactical & strategic level

I'm not sure why you disregard Shogun here. IMO, that's the hardest of the three core TW games. I can start up the game on normal and lose pretty easy - either in the tense early turns or later on encountering the late game "Hojo horde" effect. I'm not saying that's because STW has the best AI, but it suggests the AI did not hamstring the game.

I guess I'm mainly talking about the strategic level here. I believe the STW strategic AI cheats more outrageously than MTWs (ie turns are supposed to be simultaneous but the AI acts after "seeing" your move). I am glad that was toned down in MTW (and is irrelevant in RTW) as it was not fun.

In terms of the battles, I can't recall STW battles being significantly less challenging or easier to exploit than MTWs[1]. And as you indicate about modding MTW, the AI in STW tended to come to the battle with a better mix of troops than in vanilla MTW.

[1]On reflection, I guess the main big problem with STW on the battlefield was the suicidal Daimyo's. Often a few decisive early battles would eliminate several key factions and leave you facing a mass of rebel provinces. I think MTW was better in that regard (less suicidal? more heirs?).

I think STW was more challenging, but MTW provided the challenge in a more fun way.

Doug-Thompson
07-13-2006, 14:56
I think they are in danger of loosing the people with a high historical interest and a love of strategy games. These are the people from which the complaints are coming.


We don't disagree, Puzz3D. As I said later, there was a:


core decision to make the game "more exciting" at the expense of logic.

I'd argue that this was a very bad decision, but that's different from calling CA designers idiots. I don't think they're insensitive to consumer demand. I think they tried to attract new customers — admittedly, at the expense of their steady ones — and have paid a big price when the fan base revolted. I think that's part of the reason we're seeing the Medieval period again.

CA's trying to regain some loyalty, "going back to the roots," so to speak, which I appreciate and consider a good sign.

I hope the AI's much better, also. Nobody on the forum has complained more than I have of the graphics tail wagging the gameplay dog.

Doug-Thompson
07-13-2006, 15:20
CA, you can have your fancy graphics, just make a GAME... not an interactive 3D animation... 'cause at the end of the day, whatever you choose, you will still reap huge amounts of money... but end up satisfying no one if you don't heed calls for better gameplay. The shooter/clickfest market you're trying to tap will buy your product, take a look at the pretty scenery, then lose interest. Or... they could become one of us, loyal fans and veterans who feel that they've been had. :no: :shame:


:2thumbsup:

Puzz3D
07-13-2006, 15:44
CA's trying to regain some loyalty, "going back to the roots," so to speak, which I appreciate and consider a good sign.
You really believe that when you can see a cannon in M2TW, that can't rotate, fire and kill every man in a unit of mounted knights as though they just got hit by vulcan cannon fire? This in a game that purportedly models an era where large cannon were not used against infantry.



Nobody on the forum has complained more than I have of the graphics tail wagging the gameplay dog.
Well that certainly doesn't appear to have changed in M2TW. All you ever hear is the generic term "improved" with regard to the AI, and nothing about the missing features in the battle engine; features that were in the STW/MTW battle engine. Yes the programmers will try to fixup the gameplay in the first patch, but they can only overcome design decisions to a some extent. For example, we already know they won't change movement speed significantly after the game is released because it involves too much work. What M2TW has going for it is that it's a sequel based on the RTW engine, so it's likely that M2TW will not have the staggering number of bugs that RTW had which prevented the RTW v1.2 patch from addressing gameplay issues. Maybe after another two years of effort, modders will have fixed up M2TW so that it plays well, but most likely problems will remain. In the case of RTW/BI, there are strategic and tactical gameplay issues that no modder can address which adversly affect every mod that's been made for RTW/BI.

Orda Khan
07-13-2006, 15:52
I don't want TW to become just another PC game among thousands of others. No matter how good or bad the AI has been throughout the series, when it comes to the battlefield there is only one way to find a challenge.....MP. For this reason I lose interest very quickly with SP. IMO that is not good enough, the SP game needs a challenging AI desperately and I hope this will be what transpires with MTW II

......Orda

Doug-Thompson
07-13-2006, 19:55
You really believe that when you can see a cannon in M2TW, that can't rotate, fire and kill every man in a unit of mounted knights as though they just got hit by vulcan cannon fire? This in a game that purportedly models an era where large cannon were not used against infantry.

No, Puzz3D, because I don't watch game videos and certainly don't make judgements of a game based upon them. I don't pay much attention to previews, either, but sometimes read interviews with developers.

Frankly, I usually wait until I have a game play, then go to the forum. What I'm reading these days, though, isn't hard information. It's understandable fear that MTW2 will really be RTW2, only worse.

Satyr
07-14-2006, 20:02
I have been playing GalCiv2 for a couple months now and I must say that the AI has improved immensely and it is nice to see a developer that cares. The newest version made the game much more of a challenge and that is always a good thing. Even though there is no tactical battling in the game (there will be in GalCiv3, btw) the superior gameplay in the strategic portion provides a more than satisfactory experience.

Now if only CA can take note of how well GalCiv2 is selling and realize that it is entirely based on the strength of it's AI, it might bode well for M2TW. I hope all developers take note that a good AI does sell games.

Ok, I will wait for M2TW to hit the computers of my fellow gamers before I make a judgement. But if the consensus is that the AI still sucks, CA will not get my $50 and we will both be poorer because of it.

BeeSting
07-14-2006, 22:05
Ok, I will wait for M2TW to hit the computers of my fellow gamers before I make a judgement. But if the consensus is that the AI still sucks, CA will not get my $50 and we will both be poorer because of it.

I didn't buy any of their expansions for RTW after visiting the forums. Mainly becuase the AI is still stupid and I know i will get sick of the game after 2 to 3 days.

Tellos Athenaios
07-14-2006, 23:21
I suppose CA could create a really competent or 'killer' AI, but there are - in my opinion - a few simple reasons why they don't:
1 Users;
2 Profit;
3 Time.

And in that order.

Most users will be either relatively new to the game or won't have a real powerfull system on which they run the game. I mean the more AI you put in the game the more it will depend on users with a powerfull system since the game will ask more from it. Not everyone is going to buy a new computer just for one game... Henceforth CA will have to either limit the number of aspects which effectively will limit the number of troops the AI can use or CA will leave the AI relatively straightforward as it is. Clearly there will be improved AI in the new TW versions, but it won't be 'killer' AI - it won't be anywhere near human way of handling tactics.

Profit is simpler to explain: what does attract new customers? Yep, graphics. AI is a nice bonus, and will secure the more experienced lot. But still a fair amount of the veterans will still buy the new game - since they too love the graphics part. (Check the RES GESTAE mod subforum in case you don't believe me...)

Time also limits the amount of work that can be done on AI. CA has to make a certain amount of profit within a certain time limit. They are an commercial enterprise, we all know this very well. This also means you've got to balance expenditure and income. Henceforth you can't just take all the time you want to create your game as you will need income to do so, income which mostly depends on the sales of the new game. Only new releases are really in the big time - this is why most games are shortly followed up by an expansion. This has little to do with 'now we have extra time, the first was simply a sort of demo version thingy' but rather more with 'gamers are going to want more of the first, and new gamers will buy the first because of this expansion pack thingy'. This is also why RTW needed such a large amount of patches, and why the were such ridiculous flaws in the AI such as missile unit's charging melee ones.

Aquitaine
07-15-2006, 00:32
Hi everybody,

I've been around a while but mostly as a lurker. This is something I've spent some time on, and I have perhaps a unique (if not altogether insightful) perspective on it.

I am the lead designer of FreeOrion (www.freeorion.org). One of our models for 'successful' games we'd like to, uh, imitate (read: borrow/steal) things from is Total War. Obviously, we're in space, so it's not a clean transition, but a lot of the issues are the same. Strategic engine combined with tactical engine. We are currently in the process of designing the tactical engine, and starting work on the AI for both that and the strategic engine.

From the very beginning of our project, we had people clamoring for AI brilliance. They said what a lot of you are saying. We don't care about graphics (as much), just give us a challenging game! And a lot of us were in that camp as well.

The simple fact of the matter is that AI is the single most difficult aspect of game design, and the biggest time and money sink in any sufficiently large project. Our expectations of game AI are pretty high, and for good reason. Our games have come so tremendously far over the last fifteen years in almost every category, we want AI to have made similar progress. We even say 'stop spending so much on graphics, just give us AI' because of the huge discrepancy.

The fact of the matter is that, when we develop an AI capable of beating really, really good humans at Chess, it makes world news. And Chess is very structured, such that almost every operation can be reduced to math, proabability, and branch prediction. To say that Total War or any other modern strategy game isn't even close to that level of structure would be a huge understatement. We just don't have the level of understanding in this industry yet to reliably produce intelligent AI that is really satisfactory. The route most developers have taken amounts to an enormous list of weighted scripts that can be reduced to 'if this happens, do one of the following things, and if nothing happens, fall back to this behavior'. To my knowledge, only the GalCiv developers have taken the alternative route and tried to reduce all of their AI operations to something a CPU can just sit there and churn away at, such that the more CPU you throw at it, the better it is. That's something they've been after for a long time, and it's admirable - but it's also very difficult, and they have yet to demonstrate it on any kind of realtime basis, since the whole idea behind GalCiv's AI is that it ponders while you ponder. Battle AI can't do that.

All of that aside, I'm not trying to apologize for CA. None of this matters to the consumer. You and I want an awesome AI and we want it now, and we can recognize when it's broken. Doesn't matter why. But it is important to understand the problem that we (and they) are wrestling with. Fishpants already informed us of what should have been painfully obvious: the graphics folks aren't the same people as the AI programmers and it is never an either/or consideration. Like any company, CA has to follow certain business practices to remain afloat: they've got to generate excitement about their product, generate the product, and sell the product. The most efficient way to generate excitement is to release cool-looking screenshots and do interviews. Demos are always a dicey prospect, since every minute they spend putting together a demo is one less minute spent working on the final product -- and that IS an either/or proposition. Announce a demo and you've just lost dozens, if not hundreds of collective hours of development on your product. So they can't just 'show off the AI' at the drop of a hat, even if they wanted to.

Frankly, the only way AI will ever improve is if we continue to support companies for whom it is in their best interest to improve it. That means buying GalCiv and buying M2TW and trying to figure out how to make them better - something these forums do quite well. But it's foolish to suggest that CA doesn't care about AI, or that it's not a priority for them. It's just a fact that AI is behind every other aspect of gaming today, industry-wide, though there are several noteworthy exceptions that make the future look pretty exciting for everyone involved. I wouldn't be surprised if, one day, we end up buying AI-processing add-on cards just like we have GPUs today. Because truly competitive AI will require a fundamental shift in the way we think about it and in the way it's done in most mainstream products today.

Cheers,
Aq

King Bob VI
07-15-2006, 00:57
You make some good points there, but I don't think most TW fans are asking for almost human AI. Most of us would be extremely pleased to see M2TW have a decent level of AI competence that was shown in Medieval and Shogun, which shouldn't be impossible.

Doug-Thompson
07-15-2006, 03:11
That was great, Aquitaine. Thank you for coming forward.


The fact of the matter is that, when we develop an AI capable of beating really, really good humans at Chess, it makes world news. And Chess is very structured, such that almost every operation can be reduced to math, probability, and branch prediction. To say that Total War or any other modern strategy game isn't even close to that level of structure would be a huge understatement.

Consider the RTW strategy map in that context.

The Medieval: Total War map had provinces, which might be bordered by three or four other provinces. Each army had -- at most -- four or five options for movement, including the option of sitting still.

A stack in Rome Total war has literally hundreds of possible destination points to choose from.

To a layman, it appears the AI faces a vastly greater challenge on the RTW map than the MTW 1 map.


We just don't have the level of understanding in this industry yet to reliably produce intelligent AI that is really satisfactory. The route most developers have taken amounts to an enormous list of weighted scripts that can be reduced to 'if this happens, do one of the following things, and if nothing happens, fall back to this behavior'. To my knowledge, only the GalCiv developers have taken the alternative route and tried to reduce all of their AI operations to something a CPU can just sit there and churn away at, such that the more CPU you throw at it, the better it is. That's something they've been after for a long time, and it's admirable - but it's also very difficult, and they have yet to demonstrate it on any kind of realtime basis, since the whole idea behind GalCiv's AI is that it ponders while you ponder. Battle AI can't do that.

So, to give the whiners some credit, it appears you could do that with the strategy AI. It appears that each faction flashes through its options after the human player completes all his moves. If the AI allowed, say, twice as much time to sift through its options, it could be engineered to “ponder.”

But, as you point out, that would be a very expensive process.

SpencerH
07-15-2006, 13:32
Unit movement on the strategic map is not as complex as it seems though because of the various choke points and the fixed positions of the cities. Unit movement is much more complex in CIV and that has always been good but improves with every iteration of the game (it should be noted though that the calcs for AI unit movement are by far the most CPU intensive aspect of the game).

Crian
07-15-2006, 13:55
Thank you for that post, Acquitaine, I'm sure we all appreciated that. :2thumbsup:

Most of us do understand that CA is indeed a commercial company with bills to pay and shareholders to satisfy, and the fact that profit is and always will be the number one priority. We accept that, and we appreciate what CA has done as far as Total War is concerned. No other game can come close to the kind of experience Total War offers, and the direction that the company chose for the series is, IMHO, a great one. :2thumbsup:

While the majority of us aren't programmers, we at least understand (despite our limited knowledge) that making an AI that can think as good as a human is, at least for now, impossible, and the best that programmers can do is to tell the computer to "If this is A, then do B, or else do C", something to that extent. I daresay that what most of us are really asking for is not an AI that can do wonders on the battlefield... no, I'm not asking for that. It's too much to ask and we know it.

What we want is an AI that seems great at least at face value... one that doesn't frustrate you because it barely put up a fight... one that doesn't make you pull out your hair because it did something that you'd NEVER do in a million years (ie. charge your general alone into a mass of formed infantry; make your phalanx turn 180 degrees 2 meters in front of your troops, etc). At least give the AI certain absolute or near absolute DO'S and DONT'S so that it doesn't do stupid things... now I feel that this, at the very least, can be done. You might think that it might make the AI predictable and easier to beat... but I don't think so. Example, you could tell the AI to "only send missile troops to melee if: out of ammo, not the only unit engaging the target unit in melee"... something like that. That's what most humans would normally do, right? I'll be very happy with that because at least it looked as if the AI was actually thinking.

I don't want unbeatable AI (that's not fun :no:), I just want it to seem to be smarter, I hope that that can be done. :2thumbsup:

Again, thank you for your input. I do believe that the future of Total War and PC single player strategy gaming AI is bright, it'll all be a matter of time. Good luck to you and FreeOrion. :2thumbsup: :2thumbsup:

Aquitaine
07-15-2006, 14:08
I'm not an AI programmer, but I don't think you can reduce RTW's strategic map complexity just because of choke points. The choke points were probably put there (among other reasons) to reduce the load on the AI, even if only a little bit, but that doesn't do a lot to alleviate the number of factors involved in how a human moves an army on the strategic map:

THE TOTAL WAR AI, by Aquitaine

- MTW: Rebels. Fight!
- RTW: Do I ignore those rebels, or move this small stack out of a city to deal with it? Do I have enough intelligence on the rebels to know if my small stack can handle them? Or do I wait around for a returning army to clean them up on the way back? Can I weigh appropriately the danger of exposing my city versus the limited intelligence I have on neighboring enemies?
- MTW: I have a navy, I guess. That means I can make this trip in two turns instead of seven. Hooray linear pathfinding!
- RTW: I have a navy. Wtf do I do with it?
- MTW: I want to attack the Uesugi. I border them in exactly two places, so I should look at the troop counts in both places and then attack. Maybe I'll weigh the difficulty of assaulting the terrain in each of those provinces, but maybe I won't.
- RTW: I want to attack the Gauls. Setting aside for a moment that my grandmother could attack the Gauls without any help, do I head directly for the nearest city, or do I spy a large Gaul stack and move to ambush them near this choke point? How many turns can I safely lay siege to their city without exposing my flanks?
- MTW Diplomacy: I hate everyone. Hey, let's have an 'alliance.' Ha, ha. No, really.
- RTW Diplomacy: I have a migraine trying to figure out how I feel about you. Give me ten thousand florins or ducats or something to make me feel better.

...I could go on. We aren't talking a little bit more complex. We're talking an order of magnitude. But again, this doesn't forgive obvious AI blunders. It's one thing to go 'well, it's hard! sob. poor CA.' It's another to go 'when the player can spot an obvious flaw, it ruins immersion, so we should complain about that...while keeping in mind that it's unlikely any other developer would have done better.'

Aquitaine
07-15-2006, 14:14
I'm not really talking about a human-level AI. I'm just talking about an AI that can give a human a decent challenge. Not the Deep Blue of Total War AI.

I wouldn't even know where to begin on a battle AI. Even on a strategic map, the number of decisions and variables is pretty significant but can, ultimately, still be weighed. On a battle map, any decent TW player is considering and rejecting a huge number of decisions every second. To you and me, one glance at the screen tells us more about the situation than the computer can figure out just by comparing x,y, and z values of the general versus the values of the phalanx. In one sense, there is so much information available to the AI on the battle map that I don't have any idea how it prioritizes. It would be very easy for the AI to simply hold back the general nearly all of the time, just as it's easy for it to aggressively use the General. It's exrcruciatingly difficult for it to know exactly when to commit the General, whereas, for you and me, that moment is often the turning point of the battle -- the 'hold it, hoooooooold it .... now!' The AI probably knows, okay, the general is over here, and the general's unit feels this way about infantry and this way about spears and this way about cavalry, so we know we probably would have better luck charging this guy than this guy ... maybe we should do it now?

When we look at the screen, we can do almost instantly what the AI is largely unable to do at all - pattern recognition and intuitive planning. Being able to predict the exact point of weakness in your line and knowing not only where, but when to reinforce it, is just something we do a lot better than a computer can.

But look on the bright side. At least we won't be conquered by robots any time soon.

Hepcat
07-15-2006, 15:32
AI can begin to form patterns is the problem. Some of my friends found that when fighting me on LAN. They had 'perfect' strategies which always worked against AI, but when facing a human player you are never sure of how they will react.

Even when modding teams revise the AI there are still patterns and logic to the processing. It may not become apparent for a while but eventually you will find a sure way of defeating the AI. And there is also making the AI fun to fight, not ultimate.

There is a thin line between a push-over and an un-beatable enemy. It will be almost impossible to get it in the middle, because all players are different and play differently.

Azi Tohak
07-16-2006, 00:27
I don't know how anyone could make an unbeatable AI, that didn't cheat. As in the Hojo horde, or a massive morale bonus. I would love to see an AI that would try to flank me, or attack in echelon, or maybe just attack with a solid wall of infantry. I think it is pathetic how the AI uses Phalaxes in RTW. They should forever stay a wall of points. Wheel, advance, about-face... that is all they need to do. But they don't. One unit goes here, another crosses its path, and now you have a massive jumble of pikes that a swordsmen devour.

The strategic AI... now THAT would be the fun part. I understand how that one is appalling complicated. But... I suppose I wouldn't mind a little bit of cheating in terms of intel for the computer. But none of this in debt crap! Darned Hojos.

Azi

AquaLurker
07-16-2006, 04:07
I believe that most people who had enjoyed playing against tactical ai from both STW and RTW were asking the ai to be equally smart or as cunning as a human ai. All they want is an ai, that knows how to maintain a proper and decent formation when approached by or engaging the player's army.

Simply by maintaining a proper formation when defending or attacking, the ai can present a pretty challenging tactical game. Something that you can find in the previous TW series but not in RTW.

Another thing that is more aparent in RTW is the bad combination of troops type you find in AI armies during campaign games. Considering that multiple AI armies are stack in a single province on a "Risk" style campaign map in previous titles, the only time the player face a bad AI combi army were during the early stages of the campaign. I wonder if CA would make the AI form a 'proper stakes armies' before sending them to battle.

My hope would be that M2TW AI;

1.)Maintain a proper formation when defending or reacting to players movement. (and not the famous archer charge and stray pikes chasing cav routines)

2.)Form proper armies before setting off to attack anyone. (not small sucidal expeditions or peasants & 'fodder units class' stacks) I think this aspect is very important with regards to the RTW type campaign maps.

3.)Make use of the maximum range of their missile units.

A.Saturnus
07-16-2006, 21:52
It is certainly true that it is difficult to make an AI that can compete. Especially since so many things are connected. Errors on the campaign map lead to armies that easy to defeat on the battle map. Even if the battle AI was very good, it would still lose because of the piecemeal way armies are often shoved over the strategic map. However, the AI could better do with a few simple rules. Of course no simple rule can apply to all situations, but some rules are that useful that you're better off following them blindly than not following them at all. An example would be: don't charge your general into a phalanx!

Encaitar
07-17-2006, 04:43
- MTW Diplomacy: I hate everyone. Hey, let's have an 'alliance.' Ha, ha. No, really.
- RTW Diplomacy: I have a migraine trying to figure out how I feel about you. Give me ten thousand florins or ducats or something to make me feel better.

ROFL. Brilliant.

sunsmountain
07-17-2006, 11:06
To you and me, one glance at the screen tells us more about the situation than the computer can figure out just by comparing x,y, and z values of the general versus the values of the phalanx. In one sense, there is so much information available to the AI on the battle map that I don't have any idea how it prioritizes.
This is a very good point and exactly what troubles the CA programmers so much. How can you turn information into numbers, and how can you prioritize them? But there are a lot of simplifications that can and must be made, in order to make a decision. Perhaps CA is aiming too high here, leading to complex decisions which can (and do) go either way, which is always wrong in battle.


It would be very easy for the AI to simply hold back the general nearly all of the time, just as it's easy for it to aggressively use the General. It's exrcruciatingly difficult for it to know exactly when to commit the General, whereas, for you and me, that moment is often the turning point of the battle -- the 'hold it, hoooooooold it .... now!'
But the AI can figure this out, by for example looking at morale fluctuations. If a units morale drops below 0 repeatedly for 3 combat rounds, it routs. If you as an AI register that a lot of morales of your opponents units are sometimes dropping below 0 (and back up again), it's time to commit that general. Until that time, set GeneralAttackOrder = 0.

It just takes the programmer time (game development, money) & resources (like, how often do you want the AI to check this condition? we only have so much CPU cycles).


The AI probably knows, okay, the general is over here, and the general's unit feels this way about infantry and this way about spears and this way about cavalry, so we know we probably would have better luck charging this guy than this guy ... maybe we should do it now?
Yeah this is considered. Too bad cavalry still manages to charge spears head on. Again: Simplification is in order, CA are perhaps trying to implement too much of Sun Tzu's Art of War, without getting basics right, first.

That is all we ask.

Aquitaine
07-17-2006, 16:33
I think what you are asking is very reasonable. Even given a decent understanding of the complexity of the AI, it seems like CA bit off more than it could chew making the jump from MTW to RTW.

I guess it just seems like we're walking a tightrope between a generalized understanding that this is really difficult stuff and shouldn't be compared to good graphics (particularly not in the sense of a trade off) and wanting reasonable refinements of what they have already done.

I didn't post to try and shut down this argument, because it's a good one (from many of you). But a lot of people -- not necessarily in this thread -- go overboard and demand better AI and offer up graphic quality as something they're willing to sacrifice for it, and that just says to any amateur developer (to say nothing of a professional one) that you don't 'get' how their process works.

And if you want to influence their process, the most helpful thing you can do is to have a thorough understanding of it. Which means not equating an occasionally suspect AI with CA being disinterested in AI.

sunsmountain
07-17-2006, 17:21
Which means not equating an occasionally suspect AI with CA being disinterested in AI.

The troublesome thing for us is that it's difficult to test properly. You need custom battles that play out exactly the same to get reliable data. Good luck in trying that with anything more than 1 unit vs 1 unit. So we report the AI decisions on a case by case basis, which cannot usually be repeated, since battle replays cannot be saved in the campaign game, assuming we're playing single player.

But some tests have been done and they have led to some improvements in the skirmish AI. Phalanx AI and army level AI is still lacking though, and we somehow do not have ways to provice CA with meaningful data...

patient in any case until MTW2 is released... what can i dO?

Tamur
07-21-2006, 18:45
I missed it by about a week, but that was a superb post Aquitaine. Very well put and very real.

AI cards, that would be exceedlingly sweet :)