MTW II: A prediction

[parcelt]

Ok, I just read the interviews on http://www.cgonline.com/index.php?op...d=860&Itemid=1

Not much really new info in these interviews, although still nice to read. Also, I think I see enough of a pattern in the answers given in this and other interviews, to make a prediction - if I may be so bold - as to how MTWII is going to turn out on a number of key aspects:

Campaign gameplay: AWESOME. I like the new aspects of religion, castle vs city development, crusades, new agents such as merchants, etc. Overall, it is going to be a blast to play.

Campaign AI: BETTER, BUT STILL LACKING. I fear that the main disappointment here will be diplomacy. CA has consistently adressed diplomacy in interviews by referring to how easily you can find out what other factions think of you and how they will keep track of your past actions. Unfortunately, these are not so much major accomplishments but rather very basic components that are a necessary requirement for any diplomacy system to work (and their absence a first reason why RTW's was such complete crap). The main improvement that I would like to hear about is how CA has been been focussing on a new system (the current one being flawed on such a fundamental level its beyond repair) that allows for the AI to make intelligent (counter)offers in diplomatic negotiations. The issue of the two barbarian peasants - the last of their tribe - refusing to become your protectorate while facing three of your golden chevron veteran legionary stacks has been discussed over and over. It eliminates all the fun from diplomacy. I am not asking for rocket science, something on the level of CIV IV will do.

Other issues regarding campaign AI in RTW (for me the main ones being AI sieges, small useless stacks, non-use of familymembers, useless army composition) I am relatively confident will be adressed succesfully by CA.

Graphics: AWESOME. The fact that half of the interview was about the battle graphics/animations, etc, I think says it all.

Battle AI: DISAPPOINTMENT. As much as is being talked about graphics, as little attention is paid to AI development. Again, the fact that half of the interview was about the battle graphics/animations, etc, I think says it all. But also the wording used is so much weaker. This may be further illustrated by these quotes:

In the battles we’ve stripped down and totally refurbished our technology. Medieval II’s engine includes a complete upgrade of the battlefield rendering system, new methods for building and rendering cities while the combat animation system has been heavily revised.

How does this compare to, for example, the minor tweaking done to the diplomacy system? Why not 'strip down and totally refurbish' that?

Regarding battlefield AI (this is what I mean by 'weak wording', compare this to how CA talks about graphics):

We’re always striving to make improvements to the battle AI as the Total War series has developed and Medieval II will of course continue this trend.

And in general regarding the level of challenge:

We also have to balance this with developing the game for an established Total War audience and strategy fans. Medieval II will certainly present these players with an intense challenge and experience. Features like the new in-depth religious system, improved diplomacy and the option to develop settlements as cities or castles have added brand new layers of strategy that we’re sure will delight existing Total War fans.

This does not sound to me like the fundamental AI issues will be adressed, rather it seems the idea is more to introduce some new features that may distract our attention for a bit.

OVERALL: GREAT BUT STILL FRUSTRATING. I think MTWII will be a fantastic game, significantly better than RTW and overall it will be a joy to play. Still, there will remain a fair bit of unrealized potential in terms of some good new ideas that will be implemented with serious flaws, and (much worse) existing issues that are carried over from game to game in the TW series and somehow CA refuses to adress (or not sufficiently). I have a personal bet with a friend that suicide generals will still be in. These issues will continue to frustrate the TW veterans. Mmmh, I guess I'm saying not really much will change

[Puzz3D]

It seems to be a matter of priorities. I just played a campaign battle in original Shogun v1.12 where the AI was defending Hitachi province. That map has two groups of trees near the back of the defender's side on high ground. If there had been fog I wouldn't have seen this, but I attacked on a clear day and observered the AI split it's forces into two groups and send one to the trees on the left and the other to the trees on the right to set up ambushes. Since one of these groups was composed of all archers, I attacked it with two cavalry units, and I lost because my cavalry was fighting in trees. Out in the open, the cavalry would have won, but cavalry gets a combat penalty in trees.

The game crashed after I finished the battle since it's an old game running on a new machine, so I reloaded the campaign and got to play the same battle. This time the AI split its forces, but did not send a group to the trees on the left. Instead, it left a combination of archers and spearmen in the center, and sent the rest of its force to the trees on the right. I send a cavalry archer to harrass the flank of the center AI group. When I started shooting a spearmen, the AI pulled the spearmen back out of range sending it to the trees, and sent out an archer to help the archer in the center group shoot at my cavalry archer. I switched target to a second spearman, and the AI pulled that one back to the trees as well.

In a third repeat of this battle, there was heavy fog. The AI, as the defender, moved forward with its whole force and surprised me by attacking.

The AI's single bridge defense was good. In one battle, the AI repelled my first attempt to cross a bridge. After that, its ammo was exhausted, so it pulled back to high ground to defend against my second attack.

What happened to this kind of battle AI that would choose various strategies, set up ambushes and respond to ranged attack by moving a unit out of range? To be fair, the AI doesn't always move a unit out of range, and Shogun wasn't perfect. MTW/VI saw some improvements such as suicide general fixed and using cavalry to flank.

The diplomacy in STW was rudementary, but it was logical. The AI's strategic attack threshold was dependent upon it being at war, neutral or allied. I can't recall it ever attacking with numerically inferior forces. If you handed the AI costly defeats, it asked for peace. As I recall, it was sensible about accepting a peace offer or alliance. If a faction was eliminated, its territories went to its allies.

[Dooz]

Puzz, reading those mini battle reports feels like a warm fantasy. Actual tactics being used, how fun! Hope for the best, expect the worst I suppose. Oh, and parcelt pretty much nailed it I'd say, my sentiments exactly.

[Yun Dog]

yep alot of talk about astonishing graphics isnt there

I had a brief look at the MTW2 site - units - Jaguar warriors???? what is this Civ 4

I started thinking RTW looked great but then realised the game was virtually unplayable because of a terrible battle and strat AI

will MTW2 be different - alas I am resigned to being very diasppointed with what will no doubt be a graphically impressive but essetially empty shell. How many battles will you play before becoming so bored you smash the disk out of frustration -

I experienced those STW battles too puzz - where did that AI go - where did the balance go - where did the tactics go???

it seems unbelieveable that somehow it just vanished from the game
but be in no doubt these things are long long gone


a game that had years of gameplay as MTW became a 5 minute wonder in RTW - it was a completely different game - it has no relation to its predecessors

judged this way - I guess I will buy M2TW to look at those graphics and dream of what a fantastic game this couldve been - and then never play it again

perhaps one day someone will make a game that has great game play - tactics and a challenging AI - but I think that CA has stayed far from this road and those holding onto the hope it will find it again - hope in vein - the next title - maybe the next title

basically that sort of game is not a priority - a graphically impressive rts is the priority now - because they know that will sell - and given the short attention span of their target audience - who cares if the game has a 5 minute lifespan as long as its intial sales are through the roof -

then we can listen to a whole new generation of disappoited gamers - whining about this new game they got called MTW2

[DUTCH-BUDDHA]

the must came up whit new ids like multiply coop campaign or plyer vs plyer campaign whit a save option so u can ply next day. the next movement on main map or do the battels .....


DUTCH-BUDDHA

[Subedei]

This is what Pras The Reaper [CA Staff] stated in another thread. As we REALLY don´t know what the game ´ll be like exactly, I think it is an interesting post & point he makes here:

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]

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.

I disagree. We can get alot from movies concerning the AI. The german video in which a battle is played against an eastern faction still shows some issues. Of course the AI was at that time still being programmed, but when I see issues that were present in R:TW then it's hard to remain optimistic. I mean, the AI of R:TW messed up clearly and it wasn't noticed or at least not fixed. That gives me little hope that they are seeing it now, especially when they show videos with the same mistakes still present.

And with the AI messing up I mean:

Not recognizing which units are being covered and which are not. This results in units being charged by the AI multiple times while leaving others alone: you can flank the AI by doing nothing as it will expose its flank on its own.

Forming blobs of units stacked upon each other. The AI uses the army formations to move in perfect order up to a certain distance from the enemy. Once close enough you will see units breaking out of the formation (see picture 1). Even closer the missile units will picking targets. Sometimes this means moving laterally which will result in stacking. Other times the same unit is targeted by multiple units and the AI missile units will converge as they approach the target.
This should be fixed by dividing the army into task forces. For example when the AI decides to start shooting the enemy then it should collect a few missile units and put them together in a group formation. The same with infantry that is meant to charge.

And here is a little analysis of the video:

[parcelt]

Wow, that's quite an analysis, Duke John!

I think it shows once again what the main weakness of the game will most likely be: Battle AI. And given that I'm reasonably optimistic regarding most of the other aspects of the game (I just hope that I was wrong about diplomacy, please), I just realized what is going to be the make-it-or-break-it factor for this game..... autoresolve!

Because if battle AI turns out like it seems it will, the battles won't be fun for long. Eyecandy weares of quickly. Then we'll still have a great gampaign game, and with a good autoresolve there'll be no need to play battles yourself. I used to play RTW and sometimes even MTW quite a bit like that, except for real important battles.

The strategic part can definitely be a lot of fun by itself.

Of course, ideally the battles would be so much fun (which should be translated as: presenting a challenge in terms of actually facing though tactical decisions and having to do some thinking, and having the time to do so) that you would actually WANT to play them. But if that is not to be, I could still enjoy myself quite a bit with the strategic part. However, once you go that route, having to descend into battles you know will just be another little clickfest is really just an annoying disturbance. Ironically, I even got to 'like' RTW's ridiculous running speeds, kill rates, etc, because at least it would be over quickly.

Therefore, everything may end up depending on autoresolve! Given how autoresolve has worked in all past TW games though, I am not sure my new insight should make feel any better....

[Duke John]

The tactical AI doesn't need to be challenging, I just want it to start making sense.

For example; a battle on a steppe flatland map. Time after time do I see the R:TW AI try a flanking manouevre with its entire army in plain sight. Why? So I respond by just rotating my army. Then the AI admits the futility and marches towards me. However because of the lateral movement his units are now behind each other instead of in a nice formation. A senseable AI would first form up before moving. Not the R:TW AI. The consequence is that melee units arrive one by one and missile units will stack upon each other.

There are more examples which clearly show that the AI tries to be clever but loses the battle because other parts of the AI are not programmed well. And in that case I would rather see the AI not trying to be clever but just get the basics right:
- When rotating the formation wait until all units are in the correct position before moving on
- Do not frontally charge an enemy unit that is already being frontally charged by another unit.
- Try to cover all enemy infantry. Left over AI units are used either for flanking (offensive units; axemen, cavalry) or reserves (defensive units; spears, pikes, etc.).
- Only perform a flanking attack once enemy units have been pinned.
- Do not order cavalry to charge enemy missile units that are backed up by infantry (unengaged and within x metres).
- Do not try to match up units if it means that paths are being crossed.
- Never move towards single enemy units (bait). Always move towards the largest group. If there are no large groups, just wait (as this usually means that player is trying to envelop the AI).
- Form a battleline that is angled towards all threats.
- Never overlap unengaged units.

The strategic part can definitely be a lot of fun by itself.

Sure, but that part can done better by anyone who can program and has the right ideas, it isn't exactly an amazing feat of game design or programming. It is the tactical part that makes TW stand out from all the other games. The technology behind that is state of the art.

[screwtype]

I would rather see the AI not trying to be clever but just get the basics right:
- When rotating the formation wait until all units are in the correct position before moving on
- Do not frontally charge an enemy unit that is already being frontally charged by another unit.
- Try to cover all enemy infantry. Left over AI units are used either for flanking (offensive units; axemen, cavalry) or reserves (defensive units; spears, pikes, etc.).
- Only perform a flanking attack once enemy units have been pinned.
- Do not order cavalry to charge enemy missile units that are backed up by infantry (unengaged and within x metres).
- Do not try to match up units if it means that paths are being crossed.
- Never move towards single enemy units (bait). Always move towards the largest group. If there are no large groups, just wait (as this usually means that player is trying to envelop the AI).

Yes, I think if these simple principles were incorporated, plus maybe one or two more, you would have a much more challenging game. But let's face it, I think TW is moving more and more toward a kiddie clickfest than a game of tactics. At least, that's the impression I'm getting from what people are saying about the videos (I've only watched one or two because of the download times).

[Yun Dog]

Wow, that's quite an analysis, Duke John!

I think it shows once again what the main weakness of the game will most likely be: Battle AI. And given that I'm reasonably optimistic regarding most of the other aspects of the game (I just hope that I was wrong about diplomacy, please), I just realized what is going to be the make-it-or-break-it factor for this game..... autoresolve!

Because if battle AI turns out like it seems it will, the battles won't be fun for long. Eyecandy weares of quickly. Then we'll still have a great gampaign game, and with a good autoresolve there'll be no need to play battles yourself. I used to play RTW and sometimes even MTW quite a bit like that, except for real important battles.

The strategic part can definitely be a lot of fun by itself.

Of course, ideally the battles would be so much fun (which should be translated as: presenting a challenge in terms of actually facing though tactical decisions and having to do some thinking, and having the time to do so) that you would actually WANT to play them. But if that is not to be, I could still enjoy myself quite a bit with the strategic part. However, once you go that route, having to descend into battles you know will just be another little clickfest is really just an annoying disturbance. Ironically, I even got to 'like' RTW's ridiculous running speeds, kill rates, etc, because at least it would be over quickly.

Therefore, everything may end up depending on autoresolve! Given how autoresolve has worked in all past TW games though, I am not sure my new insight should make feel any better....

Im wondering whats makes you so optimistic that the strategic AI will be improved if changed at all from RTW (which lets face it - is terrible). And Auto resolve has always been complete BS - I only ever used it when I didnt care who won or lost the battle - if thats the case why have tactical battles at all - just have a HOI2 style medievil game of RISK.

If anything given the attention going into 'finishing moves' and graphics in the tactical battles - I wonder if the strategic game is not completely glossed over and forgotten about. As you say if its an 'eyecandy clickfest' theyre aiming at then having a challenging strategic AI wont be high on their priority list IMO.

so far as what ive seen the score card is

graphics: excellent

Tactical AI: fail

Strategic AI: fail

Diplomacy: fail

[econ21]

Im wondering whats makes you so optimistic that the strategic AI will be improved if changed at all from RTW (which lets face it - is terrible).

Well, it did improve from 1.0 to 1.5 and also from the RTW.exe 1.5 to the BI.exe. Playing with the BI.exe now, I find the strategic AI much less annoying than I recall from RTW 1.0. There are fewer of the small stacks wandering around and usually a large AI force has a general rather than a mere captain. The AI is making use of ships to land troops to try to take my provinces (although it should protect them in larger fleets).

[parcelt]

Im wondering whats makes you so optimistic that the strategic AI will be improved if changed at all from RTW (which lets face it - is terrible).

Well, perhaps I am overly optimistic in this respect (we'll know when the games comes out), but although I've expressed my share of criticism on RTW and some of the choices CA makes, I do try and give them the benefit of the doubt. They DID invent the TW series, after all.

Also, I feel that it's only logical that a number of major (strategic) improvements will be made, for the following reasons:
- gaming community with its endless amount of manhours has identified flaws in RTW strategic engine that CA could not identify themselves.
- a lot of these flaws should not be that difficult to solve. Indeed, as econ21 mentioned, it has been improving (somewhat) with RTW's patches. Also, a lot of the strategic issues bothering RTW have been proven NOT to require revolutionary new technology. Diplomacy is the best example I can give here, CIV IV proved without a doubt that a decent (not meaning perfect) diplomacy system is very doable these days. Well, now that I think about it this may not be the best example as all clues point toward CA being stubborn and continuing with RTW's diplomacy system (with some tweaks).... (why, o why?)
- the strategic map was new in RTW, you'd think the experience CA gained by now would have its effect on future games.
-finally, CA did make a fine strategic game befroe, with MTW (VI) (I know the comparison doesn't fully hold given the developments with RTW, but still).

[Duke John]

The problem I had with R:TW is that CA did not seem to care more than just getting everything to work. On practically all levels of the AI you see that the elements on their own work, but they rarely work together:
- AI armies do not interact on the stratmap once created (very partially fixed)
- AI armies on the tactical field do not consider the position and status of its allies.
- AI units do not consider the position or status of other units.

* with "do not" I mean not at a noticeable and meaningfull way.

These are not flaws or bugs. There are simply whole pieces of code not present. And the improvements of M2:TW over R:TW should be about including that code. And they should be easy to notice too: units not overlapping and not charging the same unit while leaving another nearby unit alone; tactical armies not attacking the player on their own but waiting for each other; full stacks moving together in groups of 4 (attacking 1 would mean the other 3 will arrive as reinforcements); strategic armies constantly recalculating their targets as opposed to fire-and-forget.

[Foytaz]

As far as I'm familiar with already released screenshots from M2TW graphics seems to be excellent so sound probably has to as well. But I can't avoid feeling that M2TW will be "incomplete" for all who experienced RTR or EB, and I'm affraid M2TW will be "poor" as Rome was before "proper" modding. In fact I expect lack of everything after RTR and EB, so there is a vast platform for modders to make M2TW complete.