Loading...
  OR  Zero-K Name:    Password:   

Forum index  > General discussion   > Help and bugs   >

Skirmish AI Questions

4 posts, 81 views
Post comment
Filter:    Player:  
sort
USrankZsword
41 hours ago
Hey team!

I have a friend who is dead set on thinking the Skirmish AI is worse when set to the players team than when set to the enemy team. Now I'm no master code diver but I did poke around some of the AI files to see if I could find anything to very this claim but came up fairly flat in anything in support of it. Did find some very fun comments while I was poking around (MORE Grizzly) but otherwise, yeah I'm not seeing anything for it. In particular they comment they feel the AI gets far more passive and doesn't produce the same unit count as the enemy.

It doesn't help I barely know where to look, but that's beside the point. I was wondering if there is some validity to their claim, as my experience I feel like the AI performs about the same on either team, which makes sense since it'd be, well, the same script pool. (assuming I am selecting the same AI for both.)

They have a wonder if it's something in the AI block specific to campaign stuff but I think campaign AI is just completely separate from main game, maybe even slightly map specific? But again I don't know where to check that. (I couldn't even find a separate AI listing in the AI's folder at least so that kind of debunks that for me but isn't necessarily impossible either if it's just stored somewhere else.

Anyway, thoughts/comments appreciated!
+0 / -0

AUrankAdminAquanim
37 hours ago
I doubt that there is an explicit handicap making the AI deliberately play worse, but I could certainly imagine a human pursuing a line of play which their allied AIs were not good at supporting.
+2 / -0


AUrankAdminGoogleFrog
34 hours ago
The AI could certainly end up worse based on player actions, but there isn't anything in there forcing it to be worse. It is also better at playing some factories than other, so there is randomness there. One test could be to give the enemy team an Inactive AI as well as the other AIs, then be completely inactive yourself.
+1 / -0
USrankZsword
32 hours ago
That about lines up with my best guess from the situation overall, and with how teams are handled in Zero-K, kind of just makes sense: a team with 3 AIs isn't going to perform much differently from a team with one AI due to the AI's near infinite APM and how resources are shared between the team, and that they'll all read a similar situation to have a similar response and all be working coherently to the same 'vibe'.

There's some minor factors that don't make that completely clean cut of course, 2 extra free factories, 2 extra commanders with their respective personal incomes and the like are of course going to affect overall progression, but there won't be any major dissonance on when to push vs. when to hold that having a player and AI might. There's also some other factors I want to ask them but are, tricky to find ways to phrase and point out without being there in the moment, like comparing Metal income and waste and the like that might factor into some of the 'weirdness'.

Thanks for the responses guys! I'll relay this over and maybe see about Frog's experiment idea. (With some extra checks and balances, like some starting factories will naturally semi-counter others and the like)
+0 / -0