Feedback on conquerors: the plight of the tactical AI

[raven63]

In my most recent HIP campaign, two Turkic conquerors appeared. First was Seljuk, with 28k troops if I remember correctly. Then about 100 years later came Sabuktagin, with about 39k troops.

Both of them had to fight the same kingdom: the Taherids, and both lost terribly in that very first war, despite out-numbering the Taherid massively, and despite the fact that the Taherids were fighting other wars simultaneously.

I did not see what happened to Seljuk, but all I remember was that he was suddenly dead, his troops gone.

I did see what happened to Sabuktagin. The AI mishandled the war so terribly. It failed to win any battles against an 8k strong Taherid army. To be clear, that's 39k against 8k.

I don't think increasing the number of event troops is a good solution. Ideally the AI should be improved to be better able to handle its wars.

The movement-lock thing could be the culprit here, since it makes bad decisions harder to reverse. The AI was never able to create numerical superiority in any of its roughly 6 battles, despite having lots of troops in the region.

Another thing that could help is to make the AI prioritize defeating large enemy armies over sieging and taking holdings. You can always siege afterwards.

 

[Lider_Picaro]

I had another AI complaint recently and and it happens that AI is something locked away from modders. The AI is what it is unless Paradox fixes it.

 

[raven63]

I had another AI complaint recently and and it happens that AI is something locked away from modders. The AI is what it is unless Paradox fixes it.

Thanks. Does HIP modify the combat mechanics in any way?

 

[Choorus]

In my most recent HIP campaign, two Turkic conquerors appeared. First was Seljuk, with 28k troops if I remember correctly. Then about 100 years later came Sabuktagin, with about 39k troops.

Both of them had to fight the same kingdom: the Taherids, and both lost terribly in that very first war, despite out-numbering the Taherid massively, and despite the fact that the Taherids were fighting other wars simultaneously.

I did not see what happened to Seljuk, but all I remember was that he was suddenly dead, his troops gone.

I did see what happened to Sabuktagin. The AI mishandled the war so terribly. It failed to win any battles against an 8k strong Taherid army. To be clear, that's 39k against 8k.

I don't think increasing the number of event troops is a good solution. Ideally the AI should be improved to be better able to handle its wars.

The movement-lock thing could be the culprit here, since it makes bad decisions harder to reverse. The AI was never able to create numerical superiority in any of its roughly 6 battles, despite having lots of troops in the region.

Another thing that could help is to make the AI prioritize defeating large enemy armies over sieging and taking holdings. You can always siege afterwards.

I've been playing many times on a fogless map just to manually handle all such idiocies, as the Hungarian king losing the rebellions he could have easily won due to priority given to sieges rather than defeating bulk of the enemy forces first. I have also seen the crusader realms rulers sitting on e.g. 9K of hard cash refusing to hire any mercenaries when losing horribly against the Jihad for their whole and only kingdom. I have seen the Byzzies mishandling wars in Eastern Anatolia and losing 5:1 against some minor Seljuk vassal. The tactical AI is now the biggest hole in the CK2 experience and the modders can't do anything about it.

 

[tahlaskerssen]

Historically, 8k vs 39k is a very winnable number.

Not that it can happen any time, and specially if they 39k is an spawned event conqueror which generally has excellent traits so i'm inclined to agree that's bad. Shouldn't lose.

HIP does modify the combat a bit, making the tactics and some special units being more impactful but.. medieval warfare was decided by

#1 - numbers 50%
#2 - capability of leaders 40%

Then came the rest.... leaders combat traits do more impact on HIP than vanilla but not enough for my liking. I'm making a personal mod modifying this.

How AI manages warfare and decides when and where to siege and engage it is hardcoded and there is nothing that can be done about it yet.

 

[Nix71782]

Solution: nominate the enemy leader's army as their capital, and their armies as their counties. Foolproof solution. Definitely nothing wrong with this.

 

[Choorus]

Solution: nominate the enemy leader's army as their capital, and their armies as their counties. Foolproof solution. Definitely nothing wrong with this.

What could possibly go wrong?

 

[IAmNotAmused]

Historically, 8k vs 39k is a very winnable number.

Not that it can happen any time, and specially if they 39k is an spawned event conqueror which generally has excellent traits so i'm inclined to agree that's bad. Shouldn't lose.

HIP does modify the combat a bit, making the tactics and some special units being more impactful but.. medieval warfare was decided by

#1 - numbers 50%
#2 - capability of leaders 40%

Then came the rest.... leaders combat traits do more impact on HIP than vanilla but not enough for my liking. I'm making a personal mod modifying this.

How AI manages warfare and decides when and where to siege and engage it is hardcoded and there is nothing that can be done about it yet.

Winning against 5:1 odds was extremely, extremely rare in history. So rare, you probably know all the battles off hand. 2:1 odds? Sure, it happens. Yarmouk might have been that high. Pliska was. 3:1? Far less common. Good armies tended to avoid such situations in the first place.

 

[tahlaskerssen]

Winning against 5:1 odds was extremely, extremely rare in history. So rare, you probably know all the battles off hand. 2:1 odds? Sure, it happens. Yarmouk might have been that high. Pliska was. 3:1? Far less common. Good armies tended to avoid such situations in the first place.

i wouldnt say extremely extremely, i would say "rare".

Right now, there is no way an army can win 5v1. Impossible, never happens. Not even 2v1, and 2:1 odds were less than extremely rare. (im talking about Ck2)

We can discuss al day how many % battles were won 5:1, 3:1, 2:1.

1%, 5%, 10%, we can discuss all day.

Ck2 has 0% chance of those things happening, and that's the issue.

 

[raven63]

I am happy to have rare occurrences like that. I definitely want to see a small army beating the odds every now and then.

The point of brining up this situation was to highlight how inept the war AI is. Even though it is quite impossible, like you said, to lose when you have such favorable odds, the AI managed to do it by breaking its army into groups that moved around with almost zero coordination.

Conclusion: it is not possible to mod the AI, so this has nothing to do with HIP, neither is it something that can be addressed by HIP in the future.

 

[Lider_Picaro]

i wouldnt say extremely extremely, i would say "rare".

Right now, there is no way an army can win 5v1. Impossible, never happens. Not even 2v1, and 2:1 odds were less than extremely rare. (im talking about Ck2)

We can discuss al day how many % battles were won 5:1, 3:1, 2:1.

1%, 5%, 10%, we can discuss all day.

Ck2 has 0% chance of those things happening, and that's the issue.

It's not far from impossible if the numbers are high enough. 5k vs 1k it's 100% a win for 5k, but I have had massive battles as Byzantium vs Mongols where in a favourable terrain and with good commanders I was able to get more than one heroic victory being outnumbered 3, 4 or 5 to 1.

I personally think combat is fine as it is.