What kind of AI do you want?

[The Spartan (Returns)]

an AI that ambushes, and is non-suicidal.

[Kralizec]

MTW AI > RTW AI .......... by a significant margin.

A better AI built from scratch sounds tempting, but may be hard to pull off.
While I recognise that building an AI that can actually challenge a human is difficult, it should at the minimum do the following: keep a rough formation, use up most of their missiles before attacking with bulk, and make sure units don't tire too much from running. These are some of the greates faults in the RTW AI.

[sunsmountain]

I can confirm that it is very hard to write an AI, and that I do not hope they'll build it from scratch. RTW AI has many free parameters, and still needs a lot of fine tuning.

So it's more likely they will combine and enhance what they have already, rather than starting anew. Besides, it'd be stupid commercially, throw away your AI code for an evolutionary game.

What I hope they'll do is add solid static starting formations that made MTW AI pretty good. Such a simple element can go a long way to creating coherence in armies and replayability. And some way to play that rock-paper-scissors game better...

[Dead Knight of the Living]

I just revisited MTW. I started a new campaign as the Egyptians. It's very weird playing MTW after over a year with RTW. But I found after going back that IMHO the AI in MTW is much better than the AI in RTW.

The AI in MTW did something I completely forgot they did when I used to play it. Whenever I have the high ground, the MTW AI almost always maneuvers to seize the high ground to one of my flanks.

I almost always have to reset my formation to compensate for these maneuvers. In RTW, the AI never did that. I'd take the high ground and all they'd do is send a cavalry unit or two to my flanks to attack from even or higher ground. But they never tried to maneuver their whole army onto even or higher ground.

It would be nice if MTW 2 mimics this behavior from MTW and improves on it a little. ONe way could be while your army is maneuvering to higher ground send a light horse archer unit or quick javelin men to harass the enemy army while the maneuver is being complete.

Since I was rarely harassed I was easily able to react to this maneuver.

[spmetla]

It just needs basics.

Infantry centre with general and a few reserves behind.
Cavalry on the wings, if weak defend against their cav if strong attack their cav
Skirmishers engage before main line engages fallback and then assist were units are losing once the main line has engaged
General is used as a reserve not as attacker

And then a few terrain changes for the AI, deny that hil, keep infantry in those woods, put archers on that high ground.

[caravel]

Building Ai is a complex thing, so i've been told. However, if CA is truely spending all this time polishing mtw2, i'd ask they spend this time on the AI. it needs a huge overhaul; anything better than an RTW AI will make me happy. and i'm willing to wait longer just for a challenging game[/I]

My thoughts exactly. CA seem to have all the time in the world to spend on the eye candy, yet the AI, is "far too complex" according to some of their fans. I am not an expert on AI programming by any standard, but it seems to me that the AI in TW games could be improved vastly. The campaign map AI and diplomacy was downright stupid in all games to date. The battlemap AI is now worse than it was in MTW and STW. None of this has occurred because AI programming is this unstable and unknown science that some of the fans would have you believe, it has come to this because CA have dumbed down their games to suit a younger audience. Younger ex RTS players don't understand the effects of flanking, morale, weather, fatigue, height advantage... so CA made these aspects less important. How?

1) Flanking

In STW/MTW it was very important. In RTW not so much, because you don't need to do it as much due to the overpowered turbocav (despite roman cavalry of the period not being that good) and the fact that flanking didn't really have the same demoralising effect on the enemy.

2) Weather

In STW/MTW, wet bowstrings, climate etc all effected a battle. Fog effected visibility as did sandstorms. In RTW none of this was a real factor.

3) Fatigue

STW/MTW, your units get tired after chasing about and slugging it out for long periods, they tire faster in the hotter climes and are much more easily defeated when exhausted. In RTW this wasn't really a factor, the effects are hardly noticable.

4) Height advantage

with RTW's smaller flatter maps, this wasn't much of an issue.

This is why AI is not an issue for CA. Eye candy sells games to the main target audience which are kids and are mostly not registered on these boards, AI doesn't. Kids want to charge their romans or knights at the little barbarian men and cut them down and watch them run, the don't want to deploy and maneouvre their troops using real tactics and a strategy.

[IceTorque]

Yes, the mods have no effect on the AI ability which is awful to be honest. I would like an AI that resembled a good MP vet

.........Orda

No way, simple tweaks to the gameplay settings have a huge effect on the AI. As for the vanilla settings, they are atrocious and I wonder how they could release the game in such a pathetic state.

e.g. Simply changing the unit formations to eight deep instead of four deep, achieves two big improvements.
1) It allows the AI to move it's battle line more cohesively, with less stop then reform, then move, then stop etc etc.
2) It gives the AI a much thicker and stronger battle line.

Also, simple tweaks to unit stats effects what the AI will recruit, and deleting the peasant units and giving the remaing units closer stats, makes it easy to get consistently strong AI armies that offer reasonably challenging battles.

As for the engine problems of 'drunken soldier physics', skirmishers that don't skirmish, command delay, pathfinding, friendly fire, unit stacking, the massive slowdown caused by unit stacking, Strat map: countless small and time wasting battles, instead of the steady build up of forces before the large and usually decisive battle, as happened with the risk style map. This is what I dislike most about RTW/BI, and I will be reading these forums to see if these problems have been fixed before I buy MTW II.

-IceTorque

[econ21]

1) Flanking

In STW/MTW it was very important. In RTW not so much, because you don't need to do it as much due to the overpowered turbocav (despite roman cavalry of the period not being that good) and the fact that flanking didn't really have the same demoralising effect on the enemy.

2) Weather

In STW/MTW, wet bowstrings, climate etc all effected a battle. Fog effected visibility as did sandstorms. In RTW none of this was a real factor.

3) Fatigue

STW/MTW, your units get tired after chasing about and slugging it out for long periods, they tire faster in the hotter climes and are much more easily defeated when exhausted. In RTW this wasn't really a factor, the effects are hardly noticable.

4) Height advantage

with RTW's smaller flatter maps, this wasn't much of an issue.

I have to say that in my experience, almost all of these points are false as regards the RTW engine (the only one I am not sure on is weather - I haven't looked for that). I agree in some vanilla games, you may not need to worry about the above effects. But they are modelled in the engine and are available to you. Play a harder campaign, fight tougher battles and you will end up relying on this stuff.

Flanking is incredibly important. Try playing RTR where morale levels are so high, if you just have frontal engagements, it will be mutually assured destruction. I would say flanking is the tactic in RTW.

I thought from Puzz3D and others than fatigue is overdone in RTW. Set up on a map edge and typically the AI are exhausted by the time they get to you and so they flee easily.

Height is an enormous issue, in my experience. Archers on a hill, or even just javelins, greatly outrange those below and seem more lethal. I get cut to pieces going up hill. I think the advantages of meleeing downhill are still there too.

People criticising RTW really need to play mods like RTR or EB. They'll find most of the gameplay they loved in MTW is still there. (And the historical aspect is far better.) The main thing we've lost is the greater strategic challenge from the risk style campaign map, but that's a gameplay vs realism trade-off IMO.

[Furious Mental]

Height advantage should be important, but the AI should know how to deal with a player in a strong position. It should try to use archers, arquebusiers and especially artillery to draw the player down. I also think that units should be more prone to chasing after the enemy, such that the AI can use feigned retreats to pull the player's soldiers down the hill. If that fails the AI should retreat in an orderly fashion to fight another day. What it should not do (1) pointlessly attack up hill with no chance of success; or (2) string its entire army out into a disparate and vulnerable line of individual units, leaving them to be easily slaughtered by the player's cavalry!

[sunsmountain]

Flanking is incredibly important. Try playing RTR where morale levels are so high, if you just have frontal engagements, it will be mutually assured destruction. I would say flanking is the tactic in RTW.

But I want high morale levels in vanilla RTW, not a mod! Especially RTR, where besides these good morale mods, other modifications are made to all files which i do not agree with at all. I play with 2 mods, thinking about the 3rd:
1. Bug fixer, by player1
2. Completed factions mod, by jpinard and others
3. Darth formations (not the mod) by darth vader
non-invasive is the key here, and i'm still looking for a good morale & speed mod. I'm not going to modify morale because morale DEFINES a unit. Giving everybody morale 20 means peasants will rout as soon as spartan hoplites. Just to get longer battles... I prefer a sense of flavor & balance more i guess..

If you need to increase morale levels before flanking becomes important,
If marching/charging speeds needs to be changed before archers actually contribute,
If [insert your favorite lack here],
Then something is wrong with the game. Most of these things could have been addressed with tweaking, but they left it as it is, to produce faster battles. To most of us here, faster battles simply means strategy & tactics get lost, and they do. So now we need to mod, while CA caters to the WCIII boys. Who bought STW and MTW? We did, not them. We deserve more than this, although we are greatly outnumbered....

I thought from Puzz3D and others than fatigue is overdone in RTW. Set up on a map edge and typically the AI are exhausted by the time they get to you and so they flee easily.

I agree, but the penalties to attack and defense don't matter if morale levels are too low. The initial charge determines everything anyway.

Height is an enormous issue, in my experience. Archers on a hill, or even just javelins, greatly outrange those below and seem more lethal. I get cut to pieces going up hill. I think the advantages of meleeing downhill are still there too.

I agree with the increased range, but not with the melee advantages if morale levels are too low. Those bonuses are there, but too small to have a noticeable effect. With cav or even infantry charging at you with the speeds that they do, you're lucky if you can get 2 volleys in instead of 1.

The main thing we've lost is the greater strategic challenge from the risk style campaign map, but that's a gameplay vs realism trade-off IMO.

Which also shows, IMO, that the campaign map AI needs to be better. Not attacking me because I built a fort somewhere? Come on...

[Kralizec]

I thought from Puzz3D and others than fatigue is overdone in RTW. Set up on a map edge and typically the AI are exhausted by the time they get to you and so they flee easily.

It's just that the AI will mindlessly run his units towards you, and by the time they'll reach you they're off course tired. MTW AI never did that.
I think it might be related to a quirk that occurs when you send multiple units to march somewhere. If you just click on the destination, some will run and others won't. It doesn't occur with click-and-drag commands.

[Alim]

What would be ideal is have the AI moddable. In many RTS games AI runs off scripts and the scripts are user-editable. I am sure if the modders would be able to take the matters into their own hands they would produce a far better AI then the RTW one. However, I realise that this is not going to happen and so the next best thing would be to go back to the MTW1 battlefield AI (and ideally improve it). MTW AI was far from perfect and often was not especially smart but at least I rarely remember it doing completely insane things, which RTW AI does on a regular basis.

[grinningman]

What would be ideal is have the AI moddable. In many RTS games AI runs off scripts and the scripts are user-editable. I am sure if the modders would be able to take the matters into their own hands they would produce a far better AI then the RTW one. However, I realise that this is not going to happen and so the next best thing would be to go back to the MTW1 battlefield AI (and ideally improve it). MTW AI was far from perfect and often was not especially smart but at least I rarely remember it doing completely insane things, which RTW AI does on a regular basis.

I just got RTW a few days ago, and the differences from MTW in the tactical battles is enormous. The MTW units had so much more cohesion - the RTW units just seem to break apart much more easily in combat (as well as cavalry that go from stopped to full speed in a fraction of a second, and throw infantry around like bowling pins!).

I would guess that you can't easily apply the MTW AI to RTW - the way unit combat works is too different. As you say, the ideal would be to have a strategic and tactical AI that modders can play around with: even if the MTW2 is pretty good to start with, modders will always come up with ways to improve it. Why do you say that moddable AI is not going to happen? Have Creative Assembly said it won't happen?

As for the engine problems of 'drunken soldier physics', skirmishers that don't skirmish, command delay, pathfinding, friendly fire, unit stacking, the massive slowdown caused by unit stacking, Strat map: countless small and time wasting battles, instead of the steady build up of forces before the large and usually decisive battle, as happened with the risk style map. This is what I dislike most about RTW/BI, and I will be reading these forums to see if these problems have been fixed before I buy MTW II.

You've just summarised in one neat paragraph all the things that irk me about RTW. Well, those things and the general increase in needless complexity and micromanagement.

Every time a new feature is put into the game, strategic or tactical, someone should ask "How will this affect the AI?". I don't the get impression this question was asked much during the development of RTW :(

[sunsmountain]

It's just that the AI will mindlessly run his units towards you, and by the time they'll reach you they're off course tired. MTW AI never did that.
I think it might be related to a quirk that occurs when you send multiple units to march somewhere. If you just click on the destination, some will run and others won't. It doesn't occur with click-and-drag commands.

I found tired units in RTW less often than in MTW, due to the fact that as long as you march, your units will stay fresh. The AI does indeed run mindlessly towards you sometimes, but that's because its units think they're stronger than yours on an individual basis. They probably are, but not if they acting incoherent, which they are.

Every time a new feature is put into the game, strategic or tactical, someone should ask "How will this affect the AI?". I don't the get impression this question was asked much during the development of RTW :(

Thing is, these questions were asked, but they were answered only once: First implement the new idea/feature, then we can worry about the AI. Now that the game engine is more or less finished (battlemap AI), the long delayed work can finally begin: Get that AI just right. You can't really perfect an AI until the game is finished, and i pray that CA has finally implemented all the ideas that they wanted to so they can finally get to work on this.

For this reason, i don't expect much from the campaign map AI, as we'll get new stuff like princesses, tradesmen & the pope; not looking forward to them as much as a good AI.

[Vlad The Impala]

MTW AI was far from perfect and often was not especially smart but at least I rarely remember it doing completely insane things, which RTW AI does on a regular basis.

I'm currently playing a campaign with the Turks, and in both big battles I fought this weekend, the AI repeatedly (and I mean: at least five times per battle) tried to flank me with heavy cavalry. While I had Janissary heavy Infantry protecting my flanks. -_-

[Puzz3D]

It's just that the AI will mindlessly run his units towards you, and by the time they'll reach you they're off course tired. MTW AI never did that.
I think it might be related to a quirk that occurs when you send multiple units to march somewhere. If you just click on the destination, some will run and others won't. It doesn't occur with click-and-drag commands.

You really think it's a quirk? I don't. I think it was intentionally done to speed up the gameplay. Then the designers had to change the AI so that it would still charge despite its units being weakened from excessive fatigue. So, you end up with the stupidity of the AI making frontal charges against stronger units which is something it never did in STW or MTW. Something isn't right when you get a more challenging gameplay by turning off a feature; namely fatigue.

[Lord Adherbal]

You really think it's a quirk? I don't. I think it was intentionally done to speed up the gameplay.

yeah, imagine you'd have to sit still for a minute while the enemy gets closer, or that you'd have to push the tripple speed button. The horror.

[allfathersgodi]

An AI built from scratch would be nice; with specific tactics and doctrines for factions (i.e. you wouldn't use the shield wall if you are Eastern European, Asian or Muslim factions).

Simply put an AI that is adaptable... I am sick and tired of being able to destroy an army piecemeal like William did the Anglo-Saxons at Hastings (though it should work at times). I'd also like to see the AI scaled, meaning if the enemy General has no command stars, the AI fights ineptly, but if the enemy general has full compliment of stars he kicks ass.

Moreover, aggressive AI for aggressive generals and a reserved AI for reserved generals, say Richard the Lionheart who was aggressive versus Saladin (who was much more reserved).

Simply put, I want an AI with few exploits, an AI that has a stored library of doctrine and tactics from which to draw upon, that will compare my force to his and act appropriately.

Moreover, I want an AI that can put me off balance. If I launch an attack from the campaign map, I pretty much know that the AI will not go on the offensive, and if the AI attacks me, I know the AI will be advancing.

Simply put, I want CA to stuff a general's brain inside the game somewhere...

Campaign wise, I want an AI that is not stupid. Example, I play the HRE, the French have been beaten by the English and have only one Territory left which borders me. I got two full stacks of troops bordering the French territory and the French are still facing off against the British. And then the French attack me...

[roman pleb]

An AI that doesn't ally with you and then attack you on the next turn would be nice.

[allfathersgodi]

Roman, if it is advantageous to the enemy then I do not see what is wrong with it. But I am sick and tired of seeing the Byzantines getting double-teamed by the GH and the Saracens, my HRE borders them in the Balkans, and they break the alliance and attack me, after twice sending my armies to join with the Byzantine army against the GH pushing through Southern Russia...

But, if I am fighting a war against the Poles and I am the HRE, and the French are not engaged, I do not see the problem with the Frenchies attacking me right after concluding an alliance...

[Lord Adherbal]

breaking alliances should have a big effect on your credibility though. AI factions should be much less likely to ally with you, and the relation with your current allies should deteriorate; which in turn could lead to them attacking you. The same should count for AI factions, so they shouldn't attack allies unless they can gain a great deal from it.

[econ21]

breaking alliances should have a big effect on your credibility though.

I thought it already does. At least I avoid it because some people here have reported such big effects on their relationship with other factions (in all three TW games).

[Peasant Phill]

I thought that in M2TW the AI will take the actions done against the other faction in account. (I heard this somewhere)
Like when one makes an offer that is laughable to the extent that it's an insult, the other faction won't be easily inclined to accept the next time.
Or
If you attack factions that are allied with you, no other faction will be willing to ally with you anymore. (of course the other way around should work too).

[sunsmountain]

What surprises me the most about this poll is the number of people who want an AI built from scratch, given the amount of time that would take and the amount of experience with RomeTW AI that would throw away.

Or perhaps we simply want improvements good enough to make the AI "as good as new".

[redriver]

I've only played Rome and so my selection was the last one. I'm pretty sure that's what CA will endup doin' anyhow... I just wanted to point out that RTW AI isn't too bad. I've had my original game bugged by the offiicial patch fix and had to delete everythin' and reinstall RTW from scratch. applied patch 1.3 then 1.5 and noticed that there's lotsa files cleared up in the root directory and I think the game AI was better as well. didn't get the large wall beseigin' tower bug either.
ok the new AI should be true to the time period. the romans were masters of warfare compared to most other factions of the period. barbarians didn't even have proper formations or unit types. just one big hip of randomly armed/armored men chargin' the enemy lines was pretty much the norm of the time... thus less depth/AI vs MTW IMO. I too hope this reincarnation of MTW will be more than just graphics update :)

[Ludens]

ok the new AI should be true to the time period. the romans were masters of warfare compared to most other factions of the period. barbarians didn't even have proper formations or unit types. just one big hip of randomly armed/armored men chargin' the enemy lines was pretty much the norm of the time... thus less depth/AI vs MTW IMO. I too hope this reincarnation of MTW will be more than just graphics update :)

I don't quite agree with your starting points. Unlike the Romans would want us to believe, the battle tactics of the so-called Barbarians could be quite subtle compared to the steamroller tactics Roman armies usually employed. Yes, the Roman legion allowed for more flexibility than most other armies at that time, but many Roman generals did not have the ingenuity to use this. I am not saying that the barbarians were sophisticated tacticians, but the idea that barbarian armies were just a mass of men going for an all-out charge is simply not true.

[Sabuti]

I just want the Ai to be better on both the battle and campaign maps. The Ai is what will make or break the game for me. If I hear too many bad reviews of the AI on fan sites I won't even bother buying the game.

[sunsmountain]

After the last couple of months discussing the AI, my general sentiment towards the AI is still in the same direction: Don't buy unless you know they've taken us seriously. Will they stop giving more beautiful breath taking screenshots and start talking about the really difficult stuff? Or will that be too boring according to marketing directors and managers which care primarily for 1 thing, exposure and cash.

And what about the possibility of letting the AI Programmer Team continue long after the Art Team has finished their work??? Anybody from CA like to comment (probably not, but hey, i can always ask )