Computer Difficulty and Bonus ?

[Slick]

Version 1.87:
...
12. Fixed - In Computer Versus Human games, the AI's would get negative happiness modifiers on treaties (partnership) they were forced to have.
...

I guess they are only "forced to have" a partnership when they initially meet???

[geoschmo]

Slick said:
Version 1.87:
...
12. Fixed - In Computer Versus Human games, the AI's would get negative happiness modifiers on treaties (partnership) they were forced to have.
...

I guess they are only "forced to have" a partnership when they initially meet???

I think that's just a poorly written history item. The happiness modifier is refering to the happiness of the populations on the planets. The other factor I was refering to is the AI anger which controls whether an AI will make a treaty with another empire. My understanding is there is no hard code switch that says "You must make a partnership with this AI race" rather what he's done is use the exsisting mechanics and simply give each AI bonus points to the AI anger variable. So they will tend towards partnership with AI players unless some other factors override that. In some tests I did creating a scenario once I could see from turn to turn some adjacent AI players would form a partnership, then get angry and declare war, then form a partnership again, in a continuing cycle. That's where I got the impression that it wasn't a hard code imperative, but a massaging of the anger values that was controlling their treaties with each other. So I think he doesn't literally mean "forced" in the history log. But I guess I don't know without someone asking him directly.

[Slick]

Interesting. I always thought that in computer vs. human games all the AI's were locked into a partnership.

[Q]

Slick said:
Interesting. I always thought that in computer vs. human games all the AI's were locked into a partnership.

That's what I see in my games. They may have a murderous mood towards other AI empires, but they have and maintain a partnership treaty.

[Blueentity]

The AI in SEIV does not "cheat" in the same sense as Civ, where all of the computer players start out with knowledge of the entire map, or the old Strategic Conquest games where the starting computer city received a bonus in building armies and the human player’s starting city was further away from neutral cities than the computer was. Only “Ancient Races” start with knowledge of the entire map; that race must “pay” for that trait in racial points, AI or not, and the map is subject to fog-of-war. The "Bonus", as previously mentioned, affects every AI's production and construction rates. Unless Team mode is activated and until a human reaches Mega-Evil empire status, the AI will co-operate with a human player as well or as badly as any other AI player (Again, no global "cheating").

However, it does appear that the AI “knows” your score, even if “Can view score of all players” is not selected. Since some of the AI diplomacy actions are dependant upon the score ratio, this is probably the only “cheat” where the AI has information that a human doesn’t automatically have.

As far as the team mode, running test games where I "check" by changing all AI to human from the "Players" menu and looking at logs, the AI immediately forms a partnership upon meeting another AI and immediately goes to war upon meeting a human. I can not answer if the treaties are permanent, but they appear to be holding for at least 50 turns, independent of anger level or mood.

[Fyron]

The AI seems to be able to see the cargo on a planet as well. It uses this information to determine if it would be a bad idea to attack it.

[Blueentity]

What cargo would that be? Fighters? Drones? At least in the beginning, the various AI will happily march to their death blowing up their escorts, frigates and destroyers against my mines and WP. Is this a mid or end game calculation?

[Fyron]

I think it is more of a deciding factor in which target to attack, not necessarily whether to attack or flee.