Simple changes by CA that would improve the AI

[Red Harvest]

Yes, I am sure that's right. I wasn't trying to influence BI, which must be pretty much done and dusted now, but just trying to look at the much discussed AI flaws in a constructive way. But you never know - given the success of RTW, it seems inevitable that there will be another TW game and CA may even patch BI somewhere down the line.

From most of what I've heard from CA and others about BI and the demo, I don't find it very likely that they are going to do much to really change the game. They chose a different route with RTW and its popular success seems to have sent the message that stimulating gameplay was not worth the effort. For gameplay, it would have been better if RTW had been roundly criticized and shunned. The unfortunate part of that is it can also badly damage a company...so it is lose-lose for players that want some challenge.

As I've said from when BI was first discussed, the schedule appeared far too short to address what was really wrong with the game. Perhaps CA will pull a rabbit out of the hat. I hope so.

Long term...CBR or someone said that CA already intends to use the RTW engine for the next game. Unless AI/gameplay criticisms are taken to heart and addressed, I don't see it being a game that hardcore TW players will be happy with.

[Deus ret.]

But having learned that lesson, I DON'T rely on my allies to be faithful. I garrison large armies on my borders with them, on the highways between our cities. And voila, no backstabbing.

What's the point of being allied then? The AI rarely attacks me as long as I keep decently-sized border guards, even when we're at war.

Generally it seems that 'medium' is the level at which it's best to play at. Battles run at an endurable speed and the AI behaves somewhat more realistic (I dare not say "challenging") on the campaign map....that's weird.

[gardibolt]

The main advantage was that eventually I persuaded the Parthians to give me military access so I could descend upon the Armenians from several directions and wipe them out. But it's true, the advantages are fairly marginal.

[Mongoose]

gardibolt

You still haven't answered my question. What setting are you playing on? Medium diplomacy isn't as bad as VH......

[Shaun]

any diplomacy is always answered 'no' by the AI!

[Oaty]

[Strategic AI:








Tactical AI:









(
Pretty much all the above should be easy for CA to code and would make the game much more fun. Modders can do a lot - I'm enjoying RTR v6.0 as much as I did MTW - but not everything.[/QUOTE]


(1)First of all the only difference between dificulty levels seems to be the A.I. has more money and more troops and with more experience. I'd say the A.I. is pretty decent for normal difficulty, but lacks any good improvements on very hard.

To solve the weaker army issue, small stacks are ok in the beginning phase but the A.I. starts making peicemeal of itself later on as it does not always make it a goal to get a full stack.

What the A.I. needs to do is assemble a big army in it's core cities and then head towards the enemy. Instead it sends half a stack and hopes for reenforcements in rout, wich occasionally does work in the A.I.'s favour(rarely).

(2) This probably has an easy solution with how the A.I. works. What the A.I. does is target a stack, if it gets there and does'nt attack because of odds, a second stack may come along like the odds and attack. Then after you beat those 2 off a 3rd stack or more comes alomg and dies too. Easy way around this is to see if there are nearby reenforcements that can reenforce before attacking and then attack.

(3) Ships MTW the A.I. was concerned with protecting it's coasts. RTW the A.I. tries to string out it's ships. Besides protecting thier coasts thier only purpose is to harass and blocakde the enemy. Also A.I. needs to blocakde for more than 1 turn with the only deterrent being a bigger meaner stack is nearby. Quite annoying when you see Pontus or Egypt blocking sea passages around Iberia, especially when it's a nice big stack.

(4) A coward lives to fight another day.

(5) Yes quite a shame the A.I. does'nt use a general for it's biggest battles. Perhaps the hiring of a general in BI was a fix for this.

(1) More combined efforts would definately be nice perhaps even putting it's center in guard mode so it's harder to break the A.I. into 2.

(2) quite simple there I always march phalanxes forward and hope they hit something.

(3) 2 good ones there. The tactic in itsef is not bad but needs to learn to pull back if melee units are close by. Archers are the worst on city assaults, the A.I. is in the city wreaking havoc and then suddenly decide to run around if they get hit even if I have missile superiority it is well worth it to them. They also need to sit back with missile superiority and force the human to come to them even if they are the attacker. Same goes for defending, where the better option is to attack the human instead of getting shot up.

(4) What's a suicidal general?

(5) I said it in 3

(6) quite obvious there stand back and wait for so many troops to come in then charge home.

(7) Well at least in RTW the A.I only gets to reposition itself(to a new location) once. was quite annoying in MTW to watch the A.I. run from hill to hill. Where the problem comes is that most hills have peaks that are across the red line wich I believe makes it hard for the A.I. to find the right position to defend from the highest ground possible.

[pezhetairoi]

Maybe one thing to make auto-resolve that much more realistic is to take into account the inherent strengths of some units. For example, rather than calculate hoplite attack at 7, they could give a bonus of 5-6 for its capability to form phalanx, etc. Also, some immense extra bonuses should go to ranged troops, and their chances of loss should be reduced--I have noticed that in autocalc all units lose troops more or less equally. But by right archers should lose virtually nothing unless you were getting a crushing defeat.

[HarunTaiwan]

Prediction: There will be virtually no change in the AI for B.I.

Sorry to be a pessimist, but it's what I expect from CA.

and look at all those pretty pictures of night battles and units that can swim!

ooooh! ahhhhh! I better pre-order B.I. right now before they run out!

It's so depressing.

[Duke John]

I did some testing to see what the AI was doing. Below are screenshots of my Sengoku Jidai mod and the AI faction is left alone on the island with no where else to go:

After pressing more End Turns there wasn't any building up going on anymore. Units from the settlements were immediately merged with existing armies (I gave armies plenty of movement points, enabling them to cross half the island in 1 turn and thus being able to merge with an army before said army was already gone).
I don't know how bad this is. It appears as if before saving the AI had a different set of routines (when there were still enemies) than after loading (when it was all alone).

After this test I saved, quitted, then set training times to 10 turns and reloaded. Very little merging of armies was made. Which made it become clear that the AI does not merge armies that are together more than 20 units. So you will very rarely get full stacks and plenty of stacks that range from 10/20 to 19/20. The code might look something like this: if size_armyA + size_armyB > 20 then stop, while it should be something like if size_armyA + size_armyB > 20 then size_armyA=20 and size_armyB = leftovers.

Edit: And now some praise for CA. I made a new test. Made a few turns, saved and reloaded. Then I saw some very nice manuevring:
- 2 or even 3 armies merging before attacking an enemy
- 1 army moving next to the enemy and another attacking resulting in a 2vs1. Sad part was that 6 armies did this in the same turn resulting in 3 2vs1's with the AI losing all. How much better it would have been if it had first merged into 2 larger armies before performing the pincer movement... the AI performing a pincer movement on strategic level?! There is hope for CA!
- and stretched out frontline with several (small) armies opposing each other; if only the AI also reinforced some of the armies and build some forts to strengthen the frontline it would be something much more worthy of a screenshot then some guys farting in a river.

[player1]

Hmm...

This reminds me of one of my own games, when AI has city with low public order, but maxed out troops.
It tried every turn to put more troops in city, but since it was impossible, it just had much of high stack armies around of the city. Kinda resambles screen you showed.

[Duke John]

Thanks player1, then I must remember to keep the AI's settlements happy so that it doesn't use its armies to create one big party.

[Puzz3D]

Thanks player1, then I must remember to keep the AI's settlements happy so that it doesn't use its armies to create one big party.

That's a good idea because in addition the AI is terrible when sallying from a city. The AI would is better off protecting its provinces with armies positioned outside of its cities.

[Shaun]

or, the AI wood be better off going back to CA and getting re-programed to be smarter!

[gardibolt]

Mongoose, in answer to your question I'm playing H campaign/M Battle.

[Mongoose]

hmm....that might be it.
diplomacy works best on H But H is too easy(IMO).semi-reasonable diplomacy or a harder game?

[screwtype]

continuing with this example, the quality or type of troops is not included in any way, so a huge army of peasents could scare off a small but elite army. so then someway of rating the units in the army needs to be included (can you tell this is going to get complicated?)

Including an algorithm which measures troop quality could hardly be described as "complicated". However, I'm sure your general point that AI programming is more complicated than it might appear to a layman, is quite correct.

even if you use a existing stat such as morale (the best single indicator while not perfect)...

I disagree strongly with this idea. Morale should NEVER be included in AI calculations about whether or not to engage in battle, because morale is a hidden factor. Likewise, leader quality should not be taken into account.

All that should be taken into account in such calculations regarding comparative troop quality, is unit equipment, and the battleground, ie the things that would be visible to an enemy army. Enemy morale and leadership ratings are part of the fog of war and should remain so.

...the increase in processing of fetching, adding and comparing every army battle could be too much.

Not at all. Such calculations would not strain a CPU in the slightest. You just have a rating for each unit type in the game, and then before initiating a battle, you add up all the unit ratings and compare to the enemy total. Simple.

[screwtype]

Edit.

[screwtype]

In my experience, I will not get a large stack of my own attacked by a tiny AI one. But what I will get fairly regularly is an AI stack that is outnumbered and consists of lower quality troops nonetheless attacking me. I suspect that the AI is setting the bar a little low when deciding whether to attack.

Yes, the way it should work is that there should be different "bars" for different difficulty levels. Then the kiddies can play on Easy and get to beat up armies smaller than them all the time. The mid-level players are attacked by armies with approximate parity, and on "Hard" the AI attacks only when it has a clear advantage.

[screwtype]

From most of what I've heard from CA and others about BI and the demo, I don't find it very likely that they are going to do much to really change the game. They chose a different route with RTW and its popular success seems to have sent the message that stimulating gameplay was not worth the effort. For gameplay, it would have been better if RTW had been roundly criticized and shunned. The unfortunate part of that is it can also badly damage a company...so it is lose-lose for players that want some challenge.

As I've said from when BI was first discussed, the schedule appeared far too short to address what was really wrong with the game. Perhaps CA will pull a rabbit out of the hat. I hope so.

Long term...CBR or someone said that CA already intends to use the RTW engine for the next game. Unless AI/gameplay criticisms are taken to heart and addressed, I don't see it being a game that hardcore TW players will be happy with.

Yes I agree, I think if CA were really serious about fixing the problems which have been so widely canvassed on the various forums, it would have shown at least some signs of doing so by now. But they don't care because the game has had such an excellent reception both in terms of reviews and in terms of sales.

The TW series has gone downmarket, it's aimed at the mass kiddie market now, demanding strategy gamers will be lucky if we get the odd bone tossed in our direction from this point.

[player1]

Yes, the way it should work is that there should be different "bars" for different difficulty levels. Then the kiddies can play on Easy and get to beat up armies smaller than them all the time. The mid-level players are attacked by armies with approximate parity, and on "Hard" the AI attacks only when it has a clear advantage.

It's not so easy as it sound.
This could pretty easily lead to passive AI that doesn't attack at all, since it fears of every single player stack.

The real problem is with match-up algorithm AI uses when attacks, since it very often loses autocalced battles. And it's the computer, the same one who runs AI, who calculates results of autocaled battles.

[screwtype]

Yes I know it's not as easy as it sounds P1, I'm just suggesting a general principle by which the various difficulty levels could be further and more effectively differentiated.

I just want a game that provides me with a genuine challenge. I've noticed with all of the TW series, the AI will often only attack you with approximate parity, even when it has overwhelming force at its disposal. On hard levels, the AI should not be fighting with one hand tied behind its back. It should be giving it to you with both barrels.

[Veresov]

I would suggest that the AI be required to use generals to move units. I see too many no general armies with the general sit in city.

[pezhetairoi]

I think that is a good idea. But that would give factions an overkill of money due to there being too few armies, a too-quick faction destruction rate since their generals tend to die in battle, and makes it a tad boring. Lots of other things'll have to change also to give this one a chance.

[Red Harvest]

I think that is a good idea. But that would give factions an overkill of money due to there being too few armies, a too-quick faction destruction rate since their generals tend to die in battle, and makes it a tad boring. Lots of other things'll have to change also to give this one a chance.

Heir count is based on territory count. While you can be short or have an excess, their are various aspects of deaths, births, adoptions that are driven by the difference between the "perfect" 1.0 ratio.

It is very difficult to kill off the AI factions in battle, unless they put all their heirs in one stack...they do this at times unfortunately. The game's built in adoption system keeps adding new family members when it is short.

The AI needs some routines to end suiciding generals, true, but having a general in charge of most battles would be a positive (especially assuming they fix this.)

[Ludens]

Great work, Simon Appleton. I have only a few comments:

Yes, in reply to LordPreston's point, I don't think this change is technically difficult. Indeed, I am pretty sure there is already some kind of AI routine where it works out whether to attack based on a computation of relative strengths. In my experience, I will not get a large stack of my own attacked by a tiny AI one. But what I will get fairly regularly is an AI stack that is outnumbered and consists of lower quality troops nonetheless attacking me. I suspect that the AI is setting the bar a little low when deciding whether to attack.

Perhaps this is because higher difficulty auto-resolves favours the A.I., so the A.I. thinks it has a big advantage while in fact this only applies to auto-resolve and not to manual battles?

(7) Well at least in RTW the A.I only gets to reposition itself(to a new location) once. was quite annoying in MTW to watch the A.I. run from hill to hill.

I disagree: it added to the challenge. The downside of sensibly redeploying A.I. is that it will redeploy sensibly, even if the other location is some distance away.

[sunsmountain]

AFAIK, The programming methods used in RTW were very similar to those in MTW. In tactical combat, units use multiple If.. Then.. rules to decide what to do, which together are consistent. The campaign map AI was entirely new code.

Most of Simon's comments are 'rules' defining what NOT to do, but you can not make anything out of negatives. The AI needs to know what to DO with the unit every second, not what NOT to do.

The Ai consists of hundreds and hundreds of lines of non-trivial code, you all are making it sound like it's so simple to fix. Simply adding another If.. Then.. statement leads to chaos, compiling trouble and later crashes. In a project as large as RTW, every line of code needs to be checked and double-checked.

Now what gives you the idea that you can think up a simple rule that will even stand the critical eye of even 1 programmer, let alone the team? I'm beginning to understand why they don't study threads like these: AI programming is well hard, and far harder than any of us can imagine. In fact I doubt even one of us has experience with this at all.

I don't know what we CAN do, perhaps simple testing is the best we can do.

As a side effect, most games focus on graphics nowadays, not just because it looks pretty or sells well, but also because it's easeier to program. Graphics code behaves. AI code doesn't (as we can see). It's not really possible for them to share this with us, since it'll be copied immediately if it's good.

[econ21]

Most of Simon's comments are 'rules' defining what NOT to do, but you can not make anything out of negatives. The AI needs to know what to DO with the unit every second, not what NOT to do.

I'm fairly happy with what the AI does. I've played a lot of strategy games and know not to expect a "deep blue" kind of super-AI. Indeed, TW's AI is already up there among the best AIs I've encountered in any strategy games. Only Civ2/3, Homm3 and Imperialism2 come to mind as perhaps having better AI. Even then, in these games, there were cheesy exploits you could use to pretty much guarantee you'd win and the process of winning was rather less fun, IMO, that the sound and fury of Total War battles. If the RTW AI has enough resources and material to work with, it already gives a fun game. At the strategic level, it builds up and forms large armies that it uses. At the battle level, it can put up a decent fight.

The RTR 6.0 campaign I've just finished is an example of this. Even by the time I took the 50th province, it was still great fun. At a strategic level, I was struggling to subdue about six major factions (none of which I knocked out). At the tactical level, I was still fighting exhausting battles with full stacks of elite Ptolemic or Seleucid armies. It's all the more impressive because TW offers many more choices - that could potentially stall the AI - than most other games.

My suggestions are simple tweaks that I think would improve the AI more. We'll have to agree to disagree over whether they would be feasible or enhance the game.

[Tim]

The AI dishonoring alliances and back-stabbing is 100% true to life. It's not realistic to be able to entirely rely that your allies will be faithful. They're just not. They look out for their own advantages and rationalize any breaches of good faith they have to incur. If they sense you are weak, they will and should attack you. If you want the alliance to be strong, you have to do things to make it so. There's not a problem here.

Please, don't pander to the poor AI. This is my biggest problem with some players. Ever notice when someone complains about how the AI just stands by and does nothing while its men are being showered by arrows, (especiallyh before first patch) whether in a battle, or particuarlly during a seige? People pop up and say, "well, it's realistic to loose some men to this." Please....

There is NO reason why you cannot maintain alliances. You do not have to conquer the whole map, just 50 provinces. You can't keep alliances for (usually) a few turns if you are close by. I can at least 'role-play' and keep my end of the bargain and watch our empires grow. The alliance system DOES NOT HAVE TO BE FULL PROOF gardibolt, but it should work much better that it does now. Let's be honest, it simply is broke in its current form.

[PseRamesses]

How hard would it have been to make a fold-out scroll list with settlements to where you want to direct enslaved populations when you´re given the option to "occupy", "enslave", or "eterminate"? This way micromanaging your empires growth would have been a possibility. As for now you have to move your govenors out of every city before the battle that will generate slaves which is tediaus as h_ll.

[gardibolt]

You can't keep alliances for (usually) a few turns if you are close by.

Pish-tosh. I was directly next to Pontus in my last Julii campaign (indeed, had them surrounded on 3 sides) and kept my alliance with them going without any problems for over 40 years. Same thing for Parthia, though I only bordered them on one side.