Simple changes by CA that would improve the AI

[econ21]

This is probably an old question, but perhaps worth a new thread. What simple coding changes do people think CA could implement to improve the AI?

To start the ball rolling here are some that strike me immediately:

Strategic AI:

(1) Do not attack when outnumbered. I regularly get attacked by AI armies that are weaker than my own and there seems no strategic reason for a desperate assault.

(2) "Double team" armies. Sometimes the AI will have three full stacks and attack me with them sequentially, losing one each time. Instead, they should put two in "contact" and then attacked with the third, so that the first two could reinforce. That would be a challenge (even if only one army was on the field at a time, I would have lost my ammo by the time the second arrived).

(3) Stack fleets. The naval war is a joke - unlike MTW, the AI builds enough ships but loses now them piecemeal because they seldom stack.

(4) Keep armies out of reach of stronger armies (ie keep inferior armies out of the movement range of stronger armies). The AI in Heroes of Might and Magic III did this very well - it makes it frustrating to bring the AI to battle, but greatly increases the chalenge.

(5) Put good generals in command of large armies. Most large AI armies are led by strutting fools and mewling infants (captains), yet there is the occaisional high starred AI generals sometimes left alone in towns.

Tactical AI:

(1) Do not attack piecemeal. It is fun to fight Seleucids in RTR v6.0, as they combine phalanxes with fiercesome sword units. But what happens is that the non-phalanx stuff charges in, gets defeated and then the phalanxes arrive. If the phalanxes hit at the same time as the other stuff flanked me, I'd be in trouble.

(2) Keep phalanxes in a line and march forward into combat - stop them veering off to make piecemeal unit to unit match ups. In vanilla, a phalanx can crush most other infantry if kept in a solid wall (my German spears could overrun massed armies of hastati with virtually no loss). But the AI can't pull this off and so phalanxes become very weak units for the AI.

(3) Do not open combat by charging in missile or skirmisher units! (Really bizarre behaviour). When attacking the AI does not seem to use ranged superiority if it has it.

(4) Do not open combat by charging generals into battle (suicide Daimyos are sometimes back in RTW).

(5) Do not stand on the defence if getting shot to death. Better try to take some of the enemy with you, than just step into the position of a unit wiped out to enemy missile fire.

(6) Do not parade up and down in front of missile fire when defending wooden walls. (Again very wierd behaviour).

(7) Do not reposition to lower ground when the player tries to maneouvre you off a hill. (This refers to Puzz3Ds observation in the recent MTW AI vs RTW AI thread.

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.

[Kor Khan]

This is probably an old question, but perhaps worth a new thread. What simple coding changes do people think CA could implement to improve the AI?

To start the ball rolling here are some that strike me immediately:

Strategic AI:

(1) Do not attack when outnumbered. I regularly get attacked by AI armies that are weaker than my own and there seems no strategic reason for a desperate assault.

(2) "Double team" armies. Sometimes the AI will have three full stacks and attack me with them sequentially, losing one each time. Instead, they should put two in "contact" and then attacked with the third, so that the first two could reinforce. That would be a challenge (even if only one army was on the field at a time, I would have lost my ammo by the time the second arrived).

(3) Stack fleets. The naval war is a joke - unlike MTW, the AI builds enough ships but loses now them piecemeal because they seldom stack.

(4) Keep armies out of reach of stronger armies (ie keep inferior armies out of the movement range of stronger armies). The AI in Heroes of Might and Magic III did this very well - it makes it frustrating to bring the AI to battle, but greatly increases the chalenge.

(5) Put good generals in command of large armies. Most large AI armies are led by strutting fools and mewling infants (captains), yet there is the occaisional high starred AI generals sometimes left alone in towns.

Tactical AI:

(1) Do not attack piecemeal. It is fun to fight Seleucids in RTR v6.0, as they combine phalanxes with fiercesome sword units. But what happens is that the non-phalanx stuff charges in, gets defeated and then the phalanxes arrive. If the phalanxes hit at the same time as the other stuff flanked me, I'd be in trouble.

(2) Keep phalanxes in a line and march forward into combat - stop them veering off to make piecemeal unit to unit match ups. In vanilla, a phalanx can crush most other infantry if kept in a solid wall (my German spears could overrun massed armies of hastati with virtually no loss). But the AI can't pull this off and so phalanxes become very weak units for the AI.

(3) Do not open combat by charging in missile or skirmisher units! (Really bizarre behaviour). When attacking the AI does not seem to use ranged superiority if it has it.

(4) Do not open combat by charging generals into battle (suicide Daimyos are sometimes back in RTW).

(5) Do not stand on the defence if getting shot to death. Better try to take some of the enemy with you, than just step into the position of a unit wiped out to enemy missile fire.

(6) Do not parade up and down in front of missile fire when defending wooden walls. (Again very wierd behaviour).

(7) Do not reposition to lower ground when the player tries to maneouvre you off a hill. (This refers to Puzz3Ds observation in the recent MTW AI vs RTW AI thread.

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.

I agree with all of those, I've just got one to add:

When fighting an army with a battleline consisting entirely of phalanx unit, and your line is made out of non-phalanx units, don't engage the enemy head. Stretch your line out far, so as to hopefully envelope the phalanx, and keep a few units (make sure some of these are cavalry) in reserve to flank the enemy. Make sure that the units attacking the front of the phalanx are cheap and discardable ones. NEVER CHARGE A PHALANX HEAD ON WITH CAVALRY!

[CBR]

Yes and some of does seem to be very easy to implement...at least the outnumbered bit, double team and putting good generals in big stacks.

I have also had fun with RTR last week but the AI is very stupid sometimes and it just feels so easy to correct some of the worst stuff.


CBR

[Mongoose]

RTR is great. Sadly, the AI kinda ruins it a little.


phhyrus(sp?!), in my game as Rome, left his army and went to manage a small town. I fought 2 battles where he could have easily commanded an army but no....he would rather manage croton.

[CBR]

phhyrus(sp?!), in my game as Rome, left his army and went to manage a small town. I fought 2 battles that he could have easily commanded an army but no....he would rather manage croton.

In my campaign he stayed at Tarentum while his huge army stayed in Croton...sigh.


CBR

[Red Harvest]

Good list Simon. On the high ground thing...a big problem is not just the pivot. The AI won't slide laterally with its forces to block the player attempting to gain its flank via a parallel march.

mongoose brings up another point, while I was editing the strategic map to make it more challenging, I found that the AI inherently likes to split its forces for no apparent reason. If I put a nice stack of rebels in a city, the AI would take all but a handful out of the city and march them off to the hinterland to stand alone. The AI would be defeated in detail, even by other AI.

[magnum]

As far as the splitting units throughout a region goes, I believe it was likely the solution to how the AI would handle 'guessing' what the player would do. Unlike MTW where provinces one a single whole and entering it would cause the attacker to face all units within that province, in RTW players can feint attacks, do amphibious landings, and simply run around known AI units. Thus the AI tends to scatter its units so the player will have to engage them. You'll notice that once you've entered their territory and attacked something that the AI does tend to quickly gather all the units together to fight you. The AIs actions seem to be a simply solution to a complex problem. (Simple, but definately not strategic)

Overall though I agree with the majority of items on the list.

[jimmyM]

Hmmm... surely a better idea would be to put small delaying/scouting forces in the province+the majority in the city? speculation, speculation, ho hum...

(2) Keep phalanxes in a line and march forward into combat - stop them veering off to make piecemeal unit to unit match ups.

can't overemphasise this. the way the A.I thinks about these guys needs to change...I can pull a.i phalanxes out of formation into a big amorphous blob in the centre of the field with skirmishers they try to chase, then charge in and finish them quickly...but they really should keep in blocks or they're dead and useless - even if it got to the point of the A.I going "these guys are in battle line, they're not deviating 'till they contact the enemy line/general, etc." of course the problem being how do you judge what the line will need to do once its been formed from constituent phalanxes, thtere we go...the whole idea of carefully controlled setups gets derailed by the over-detailed unit-unit matchups the AI does...

[AntiochusIII]

Not exactly AI but I think diplomacy also counts.

A simple code change (I presume) could help a lot if:

1. AI honours the alliance they have with you. If changing their behavior to actually help you directly is too complicated for BI right now then at least make them more honourable in keeping their peace. It shouldn't be too hard.

Reason: players, even (or especially) casual players like it when AI act sensibly and keep their alliance with them. I know it's Total War but the game's diplomacy that actually works would please many people indeed, hardcore fans or not. People tend to keep their alliances with a trustworthy AI for a long time or even for the rest of the game if the CA coders fear for their lives and want them to have a chance

2. Protectorates honours their "alliance" with you first even if you are the aggressor against their allies. A simple code change, I believe.

Reason: It angers many people when they lost their protectorates just because they attack that protectorate's allies. It's absurd and often forces you into some irritating diplomatic position ("I forced [faction] to be my protectorate but why the **** is this subject of mine betray me for their ally? I'm their master!" - situation happens often)

[gardibolt]

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.

I've not had enough experience with protectorates to address the second point.

[manbaps]

But when the ai offers an alliance/ceasfire then attacks the very next turn whats the point? Backstabbing is fair enough but one turn is hardly enough time to lure your ally into a false sense of security.

As for protectorites there should something coded like there not able to attack for a certain number of turns on top of automatically siding with you. The whole point of them is that the protectorite should be in a position where is will lose so the ai shouldnt break this alliance unless it thinks it can win, breaking the treaty one turn after its made is retarded.

[AntiochusIII]

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.

I've not had enough experience with protectorates to address the second point.

But the AI will ALWAYS break the alliance with you whether it's sensible or not. They often break it within..say..1 turn. This should be addressed. That they backstab is fair enough, but be sensible when do so, and I think it should be dependent on the faction characteristics as well, but, of course, it would take a bit more coding for that.

To ally with one faction means alliance...not temporary peace which means absolutely nothing different from neutral stance...

[Mongoose]

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.

I've not had enough experience with protectorates to address the second point.

No. this is incorrect and the statement shows lack of experience.

I Do *not* have a problem with the AI attacking if you are *weak*. The issue here is that the AI attacks when you are very *Strong*. I have 20 provinces. you have 1. I have a huge army and you have two units of peasants. i have given you 1000's of denarii. why are you attacking me?



The problem is that you cannot rely on them *1%* of the time. The AI will attack no matter how many provinces you have and no matter how much money you give them. A big part of the problem is that the AI will often do this 100% to their *disadvantage*

Some times the AI will

1:Pay *you* for a cease fire
2:Attack

And all in the same turn!


I am sorry, but the AI attacking you no matter how strong and friendly you are, refusing a cease fire even though you haven't fought in years, and paying you for a cease fire that they are going to break on the same turn is 100% BS.

[Shaun]

i know, the AI is very random, and even the best mods are boring now because of the AI, like RTR 6, the AI is so rubbish that the RTR teams work is ruined(cos RTR is a great mod), its not the modders fault, its CAs for making crap AI. if the modders cood hange the AI, RTW wood be great, but as the AI is as clever as my cat, then RTW is just too easy. All the diplomatic stuuf os pointless, the AI isnt clever enough to know what to do!

[Seamus Fermanagh]

I Do *not* have a problem with the AI attacking if you are *weak*. The issue here is that the AI attacks when you are very *Strong*. I have 20 provinces. you have 1. I have a huge army and you have two units of peasants. i have given you 1000's of denarii. why are you attacking me?

A good point here. Moreover, it's not as though you catch one of their assassins sabotaging or trying to kill one of your dips -- that kind of subterfuge might make sense. Instead, those peasants lay siege to your lev 5 town.

As to the starter of this thread, I wonder how easy some of the proposed alterations would be to code? They seem fairly localized to me, but....

SF

[Shaun]

well, CA better do these "simple changes" in BI! these little changes cood make RTW good, by having an AI that is clever and can challenge the player, the TW series wood benifit!

[BobTheTerrible]

i know, the AI is very random, and even the best mods are boring now because of the AI, like RTR 6, the AI is so rubbish that the RTR teams work is ruined(cos RTR is a great mod), its not the modders fault, its CAs for making crap AI. if the modders cood hange the AI, RTW wood be great, but as the AI is as clever as my cat, then RTW is just too easy. All the diplomatic stuuf os pointless, the AI isnt clever enough to know what to do!

Well you have to understand, if they released the AI SDK (is that the abbreviation I'm looking for?) then people could accuse them of forcing the players to make their own AI, and being too lazy to make a good one themselves. Although I'm sure most people would want them to release it, it would be modder's paradise.

[Lord Preston]

most of the improvements do seem to be simple If statements:

If player_unit_count > ai_unit_count then
.............status = dont_attack
elseif player_unit_count =< ai_unit_count then
.............status = attack
endif

this is an example, BUT it could be more complicated and other exploits created.

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?)

even if you use a existing stat such as morale (the best single indicator while not perfect) the increase in processing of fetching, adding and comparing every army battle could be too much.


so my point is, they seem "simple" but thats how you introduce new bugs/exploits and every extra bit of code with validationand checks will increase computing time, for all the improvements listed here i think there would be a notacable increase.

Im sure the AI programmers have come up with these aims but were not able to impliment them due to not everyone having super computers.

hopefully the professionals can come up with ways to get most or at least some of the improvements into BI without making it so you can go eat dinner while it computes a turn

[Mongoose]

mongoose brings up another point, while I was editing the strategic map to make it more challenging, I found that the AI inherently likes to split its forces for no apparent reason. If I put a nice stack of rebels in a city, the AI would take all but a handful out of the city and march them off to the hinterland to stand alone. The AI would be defeated in detail, even by other AI.

Interesting. I haven't really seen it as you describe it, but in my games the AI will almost always take the 10 star commander to some God forsaken pit while it's army marches off to doom.

10 star Greek general: Remember, this battle is for our country. If we are defeated, there will be no retreat. We are even in numbers but with my amazing tactical skill we may defea...

Greek messenger: Sir, you are urgently needed at croton!

General: Why?

Greek messenger: The town has passed a new law regarding the number of names you can have for a mule. They need you to sign it.

General: I will go there right away! *gallops off at full speed*

Greek soldier: i don't think he's coming back...

Greek soldier two: What do we do now?

Greek solder one: I have the perfect plan! the army is saved!

Firse we send half the army to attack our allies in sicilly, then we send the other half to attack our allies in macedonia!

Greek soldier two: Sounds good. what about or cities in italy? won't the Romans take them?

Greek soldier one: you're right. We should do something else...**Player reloads**

Greek soldier two: What were we talking about about?!

Greek soldier one: i don't know. let's just carry out the plan.

[sapi]

Strategic/Tactical ai:

Take into account strategic (ie production capacity, economic income, overall troop numbers) rather than tactical (ie outnumbering the local fleet) when declaring war.

[Creeper525]

most of the improvements do seem to be simple If statements:

If player_unit_count > ai_unit_count then
.............status = dont_attack
elseif player_unit_count =< ai_unit_count then
.............status = attack
endif

this is an example, BUT it could be more complicated and other exploits created.

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?)

the AI should run an auto resolve of a battle before actually engageing in it, if the fake auto resolve doesnt come up in their favor, they should not attack

[LestaT]

Not exactly AI but I think diplomacy also counts.

A simple code change (I presume) could help a lot if:

1. AI honours the alliance they have with you. If changing their behavior to actually help you directly is too complicated for BI right now then at least make them more honourable in keeping their peace. It shouldn't be too hard.

Reason: players, even (or especially) casual players like it when AI act sensibly and keep their alliance with them. I know it's Total War but the game's diplomacy that actually works would please many people indeed, hardcore fans or not. People tend to keep their alliances with a trustworthy AI for a long time or even for the rest of the game if the CA coders fear for their lives and want them to have a chance

2. Protectorates honours their "alliance" with you first even if you are the aggressor against their allies. A simple code change, I believe.

Reason: It angers many people when they lost their protectorates just because they attack that protectorate's allies. It's absurd and often forces you into some irritating diplomatic position ("I forced [faction] to be my protectorate but why the **** is this subject of mine betray me for their ally? I'm their master!" - situation happens often)

Maybe a trigger that make your rival ally (which is stronger/bigger/bordering) to attack you when your faction leader dies only.

Just a thought.

[caesar44]

No. this is incorrect and the statement shows lack of experience.

I Do *not* have a problem with the AI attacking if you are *weak*. The issue here is that the AI attacks when you are very *Strong*. I have 20 provinces. you have 1. I have a huge army and you have two units of peasants. i have given you 1000's of denarii. why are you attacking me?



The problem is that you cannot rely on them *1%* of the time. The AI will attack no matter how many provinces you have and no matter how much money you give them. A big part of the problem is that the AI will often do this 100% to their *disadvantage*

Some times the AI will

1:Pay *you* for a cease fire
2:Attack

And all in the same turn!


I am sorry, but the AI attacking you no matter how strong and friendly you are, refusing a cease fire even though you haven't fought in years, and paying you for a cease fire that they are going to break on the same turn is 100% BS.

Agreed .

[econ21]

the AI should run an auto resolve of a battle before actually engageing in it, if the fake auto resolve doesnt come up in their favor, they should not attack

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.

[gardibolt]

No. this is incorrect and the statement shows lack of experience.

I Do *not* have a problem with the AI attacking if you are *weak*. The issue here is that the AI attacks when you are very *Strong*. I have 20 provinces. you have 1. I have a huge army and you have two units of peasants. i have given you 1000's of denarii. why are you attacking me?



The problem is that you cannot rely on them *1%* of the time. The AI will attack no matter how many provinces you have and no matter how much money you give them. A big part of the problem is that the AI will often do this 100% to their *disadvantage*

Some times the AI will

1:Pay *you* for a cease fire
2:Attack

And all in the same turn!


I am sorry, but the AI attacking you no matter how strong and friendly you are, refusing a cease fire even though you haven't fought in years, and paying you for a cease fire that they are going to break on the same turn is 100% BS.

Well, that certainly is different from my experience. I've cultivated an alliance as Julii with both Pontus and Scythia, and it has lasted for nearly 20 years (40 turns), even though I border both of them. I hold 83 territories and Pontus holds 4 and Scythia holds 5. Maybe you're doing something to offend your allies. I've also had an alliance with the Macedonians against the Greek Cities fall apart when I left my territories bordering Macedon too poorly garrisoned.

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.

[The Stranger]

I agree with all of those, I've just got one to add:

When fighting an army with a battleline consisting entirely of phalanx unit, and your line is made out of non-phalanx units, don't engage the enemy head. Stretch your line out far, so as to hopefully envelope the phalanx, and keep a few units (make sure some of these are cavalry) in reserve to flank the enemy. Make sure that the units attacking the front of the phalanx are cheap and discardable ones. NEVER CHARGE A PHALANX HEAD ON WITH CAVALRY!

why not it works friggin brilliant, the companions blow a hole in it and pushes the line back until its seperated in two, with only a tiny amount of casualties. now thats bizarre, how can horses that are unarmoured (i dont think armour really matters thaT much in real) push a phalanx line (phalanx pikemen) back till it seperates and suffer minimal losses. what is that?????

[Jambo]

Simon, the suggestions you've put forward as simple ways to improve the AI are spot on. However, I'd hazard a guess that if these haven't already been addressed in BI at this stage in development, then they won't be in by final release...

[Mongoose]

Well, that certainly is different from my experience. I've ivated an alliance as Julii with both Pontus and Scythia, and it has lasted for nearly 20 years (40 turns), even though I border both of them. I hold 83 territories and Pontus holds 4 and Scythia holds 5. Maybe you're doing something to offend your allies. I've also had an alliance with the Macedonians against the Greek Cities fall apart when I left my territories bordering Macedon too poorly garrisoned.

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.

Maybe the money i gave them offended them? Some factions will often refuse money as a gift for what ever reason. Is the AI so good that it has it's own "iron man" rules?


What difficulty setting do you play with? Medium diplomacy isn't quite so bad....

[econ21]

Simon, the suggestions you've put forward as simple ways to improve the AI are spot on. However, I'd hazard a guess that if these haven't already been addressed in BI at this stage in development, then they won't be in by final release...

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.

[Lord Preston]

I was just trying to make a general point that while they "seem" simple changes it could be more complicated and hard to impliment without serious performance problems.

using auto resolve to judge what they attack i serously doubt is a viable option, what if there were 3 of your armies and 3 of the AI's armies within attacking distance of each other.

It could attack each army with 1 stack of its own, of group 3 armies and attack 1..... theres 18+ possible auto resolves for it to calculate on just a small area of the map without a actual battle occuring.

im the same as everyone else, and i do expect some of these improvements listed to be included, if a enemy General charages a phalanx in BI i would be very shocked.