AI Development Discussion and Feedback

redrum

@luhhlz Agree. The AI tries to buy a variety of units and tries to categorize them into types like fodder vs defensive vs offensive vs multi-move. It then uses these categories and checks the situation around a factory like whether its threatened or how far from the front it is to balance weights across these. Then it weights the units within each category (so infantry in NWO have a very high weight in the fodder category) and randomly selects units based on these weights.

So for say a factory near the front the AI on NWO will try to buy some fodder (usually infantry) and some defensive/offensive units (elite, artillery, etc) to give more strength. For factories further from the front, you'll see it buy more tanks, mech inf, mot inf, as it wants higher move units.

Trying to determine how much of each category and how to weight units within each category is tough as maps can have very different unit sets. The above is also somewhat simplified as there are a bunch of other factors as well but those are kind of the core factors.

Black_Elk

In the enthusiasm for news of new tripleA WW1 map stuff, I've been watching the AI try to stalemate itself in that old Great War map I made with Surtur ages ago hehe. Currently skynet has bloodied itself into the 10th round. Going to just let it go tonight till it breaks itself or my memory. So far the HardAI seems to be managing fairly well for itself on either side. A dose of nostalgia from an oldy. For all its idiosyncrasies, it was fun map for the time, trying to bend the engine in early days hehe. But I haven't actually opened it up since 4.0, and now the AI is much much improved from back then.

Its been fun to watch the machine actually cobble together a vague kind of plan, compared to the nuts ass stuff it used to do with the AI when this map was first made like 10 years ago hehe. Cool to have an AI that gets the job done, pretty damn well actually, with something based on the old v2 rules.

CP are finally pushing where they can, but Allies just saw the US hop into the mix, so still a nail biter over here lol. I don't think the boys'll be home by Christmas this year. Nice work Hard AI

Great War Hard AI round 10.tsvg

Getting my kicks off for the era, watching the AI lamp itself before I crash into bed hehe. But will dream of badass 1914 PvP stuff to come for sure.

night all

redrum

@Black_Elk Yeah, I did some very limited testing on Great War with the AI a while back. It does alright as for the most part the rules of the map are very similar to revised. The only weird part is the USA maze to enter the war. But the AI eventually gets through that probably just a few turns slower than most human players.

And since the AI doesn't care about the cut outs, in some ways it has an advantage over newer players that tend to struggle with those.

Black_Elk

Yeah the various colonial boxes and naval abstractions were probably a bridge too far I think, though I still kind of like the western front box as a novel way to make the fighting in France and Belgium feel more thematic and engrossing, given the limitations of the time. One thing that the map failed to resolve, similar to many WWII maps is the overall advantage of attacking G directly via amphibious into Hanover or a Baltic push. I guess a simple solution might have been like some impassible sz to block the approach at the danish straits, but didn't really have an easy way to do it without cutting G off from the water. US entry can get pretty gamey when they arrive in force and start getting the dead drop on G amphibious. In that way it's still falls into a revised WWII trap of sorts, but still pretty Russia centric for the most part. The AI exploits that to good effect, as I imagine most humans would too, even if it takes a little longer to get there. CP AI finally handled the East after several more rounds, though the AI Allies are trying their best to push across Italy with US now. All in all the AI does its thing pretty well here on balance. What they got going with the Turks, nobodies business haha.

Great War Hard AI round 23.tsvg

redrum

@Black_Elk Yeah, especially in WWI, amphibious assaults were near impossible at any decent size scale. Even in most of the WWII maps we have, amphibious assaults are way too easy. I'm hoping to better capture this fact in the new WWI map that we are working on.

Cernel

I suppose it is well known that the AI is very bad at protecting the "combat infrastructures" (meaning infrastructures able to do something in battle), both mobile and not mobile ones, but I want to show this off as an extreme case, the AI leaving, for absolutely no reasons, a stack of 3 mammoth alone for the taking, while having several combat units that could have easily protected them.

redrum

@Cernel Mind adding the save game as well?

Cernel

@redrum Ok, I'll post it once the game ends (likely by surrender). Btw, I'm playing it in the lobby now (so, anyone can join it).

redrum

@Cernel I did a quick test with the latest pre-release setting up a similar scenario and I believe this is already fixed. I made some improvements when we were doing some Dragon War testing that I think now handles the isInfra land units better.

Cernel

@redrum Ah, that would be cool to test. Sadly, for compatibility reasons, I cannot feasibly use the prerelease in lobby. Anyways, here it is the savegame (the game just finished on round 12 with West (that is me) winning by surrender):
01_WIN.tsvg

Captain Crunch

DID AN ADMIN CHANGE ANY POSTER FORUM EMAIL NOTIFICATION SETTINGS?

I ask because I just opened up my email and see 5 emails from Triple A about this thread ... the email says I can adjust my email settings and gives the link and I saw 2 places where it will send emails so I changed it.

Was just weird I never got email notifications before (hate them) so something with this thread made me get 5 emails from these new replies but maybe the 2 places I just changed now I won't get emails

prastle

@Captain-Crunch Ya I also noticed I am receiving them again like in the past but I have not for awhile. I never changed any settings but perhaps @RoiEX did?

RoiEX

I didn't change the setting either, I had the same issue as well, but changing the setting was 10 seconds so I forgot about it

Sculpt

@redrum (or AI developer(s) -- Regarding assisting with AI for triplea.

First wanted to verify the last release is TripleA_1.9.0.0.13066_windows-64bit? And if I (us single-player/AI users of triplea) should be using, testing and commenting on the AI for 1.9.0.0.13066, or should we be using 'TripleA_1.10.14966_windows-64bit' (for instance) instead?
I received recent emails from axisandallies.org, which prompted me to d/l 1.9.0.0.13066 and play classic WWII 3rd edition, and notice what the AI was doing. So I don't want to waste anyone's time until I know which version I should comment on.

I think the AI in triplea has been excellent and a great endeavor. Thanks!

redrum

@Sculpt Responses:

1.9.0.0.13066 is the latest stable release. There have been some fixes and improvements since then which can be tested in the latest pre-release found here: https://github.com/triplea-game/triplea/releases. I recommend using the latest pre-release if you can so that you get all the latest changes but feedback on either release is still useful.
Great to hear. There are many great maps that you can download and play with TripleA. I'm glad to have AI feedback on any maps which you are playing.

Cernel

Playing "A Song of Ice and Fire" with all players assigned to HardAI the game crashed on round 16 at the start of the Combat Move phase of FreeCities.

You can find the link to download the map here:
http://tripleadev.1671093.n2.nabble.com/The-Game-of-Thrones-map-bug-tp7587435p7587606.html

You can play this map with TripleA 1.8.0.9, but only if unzipped and with a case insensitive system.

Savegame:
BugReportCernel_A Song of Ice and Fire_1.8.0.9_20190809_01.tsvg

Console:

triplea.engine.version.bin:1.8.0.9
WARNING [Triplea start thread] ProAI ->       7-MartellCombatMove: Greenstone, remainingUnitsToLoad=[HorseArcher owned by Martell]
WARNING [Triplea start thread] ProAI ->       7-MartellCombatMove: Greenstone, remainingUnitsToLoad=[HorseArcher owned by Martell]
WARNING [Triplea start thread] ProAI ->       9-MartellCombatMove: Stonehelm, remainingUnitsToLoad=[HorseArcher owned by Martell]
WARNING [Triplea start thread] ProAI ->       9-MartellCombatMove: Stonehelm, remainingUnitsToLoad=[HorseArcher owned by Martell]
WARNING [Triplea start thread] ProAI ->       11-MartellCombatMove: The Disputed Lands, remainingUnitsToLoad=[HorseArcher owned by Martell]
WARNING [Triplea start thread] ProAI ->       11-MartellCombatMove: The Disputed Lands, remainingUnitsToLoad=[HorseArcher owned by Martell]
WARNING [Triplea start thread] ProAI ->       12-MartellCombatMove: Mistwood, remainingUnitsToLoad=[HorseArcher owned by Martell]
WARNING [Triplea start thread] ProAI ->       14-MartellNonCombatMove: Mistwood, remainingUnitsToLoad=[HorseArcher owned by Martell]
WARNING [Triplea start thread] ProAI ->       14-MartellNonCombatMove: Mistwood, remainingUnitsToLoad=[HorseArcher owned by Martell]
Exception in thread "Triplea start thread" java.lang.IllegalArgumentException: Comparison method violates its general contract!
	at java.util.TimSort.mergeHi(Unknown Source)
	at java.util.TimSort.mergeAt(Unknown Source)
	at java.util.TimSort.mergeForceCollapse(Unknown Source)
	at java.util.TimSort.sort(Unknown Source)
	at java.util.Arrays.sort(Unknown Source)
	at java.util.List.sort(Unknown Source)
	at java.util.Collections.sort(Unknown Source)
	at games.strategy.triplea.ai.proAI.util.ProAttackOptionsUtils.sortUnitNeededOptionsThenAttack(ProAttackOptionsUtils.java:149)
	at games.strategy.triplea.ai.proAI.ProCombatMoveAI.tryToAttackTerritories(ProCombatMoveAI.java:1003)
	at games.strategy.triplea.ai.proAI.ProCombatMoveAI.determineTerritoriesToAttack(ProCombatMoveAI.java:335)
	at games.strategy.triplea.ai.proAI.ProCombatMoveAI.doCombatMove(ProCombatMoveAI.java:139)
	at games.strategy.triplea.ai.proAI.ProAI.move(ProAI.java:165)
	at games.strategy.triplea.ai.AbstractAI.start(AbstractAI.java:436)
	at games.strategy.engine.framework.ServerGame.waitForPlayerToFinishStep(ServerGame.java:564)
	at games.strategy.engine.framework.ServerGame.runStep(ServerGame.java:458)
	at games.strategy.engine.framework.ServerGame.startGame(ServerGame.java:281)
	at games.strategy.engine.framework.startup.launcher.LocalLauncher.launchInNewThread(LocalLauncher.java:74)
	at games.strategy.engine.framework.startup.launcher.AbstractLauncher$1.run(AbstractLauncher.java:47)
	at java.lang.Thread.run(Unknown Source)

redrum

@Cernel I'm fairly certain that issue was fixed in 1.9. Once I update the map to 1.9 then we can try to reproduce it.

ff03k64

I was just playing Middle Earth on the latest AAA version. It seemed like the hard Evil AI wasn't quite aggressive enough. Other than that, they seemed to play very well, so thanks for all the hard work!

beelee

Idk if it's been 8 weeks or not or if this is even the right thread but for some reason I just ... well

https://www.youtube.com/watch?v=dGu6v2OjW8s

Scorpions comes up next just fyi

ahh yea wrong place please move

https://forums.triplea-game.org/topic/381/favorite-playlists-songs-while-vanquishing-foes

ff03k64

Hard AI didn't seem to handle Strategic bombing very well. Was playing one of the star wars galactic war maps. First couple of times it sent several bombers in to bomb, but after I started to defend against it, it would send in 1 or 2 bombers when i had 4 fighters and AA in the area and just chewed them up. It was almost as if the AI didn't consider the fighters when it made it's calculations on if it would be worthwhile to do the bombing run.