Mali AI needs some work

[schondette]

I see this happen a lot where the Mali AI wont settle the nice giant desert biome that is suitable for the civ and instead rams its stupid face right into the back yards of other civs. Reasons like this is why i have a such a difficult time getting a game going. Funny. When I play as Mali, my desert is like 4 tiles. But whatever. Now he is super vulnerable and this game is ruined.

 

[Troy Bruckner]

It's probably because they didn't make a separate Ai for Mali so they avoid settling in deep desert just like any other Ai.

 

[MonkeyPaw]

Mali gets bonuses from desert but it also compounds their production penalty. And leaving quality city locations to other civs just because they don't fit your "theme" isn't smart gameplay either.

 

[Pfeffersack]

Mali gets bonuses from desert but it also compounds their production penalty. And leaving quality city locations to other civs just because they don't fit your "theme" isn't smart gameplay either.

Settling "normal" places is a problem for the Mali IMO - yes, desert doesn't help them with production, but thats the wrong approach. In terms of production, the Mali are kind of doomed anyway. Even in classic fertile regions they will never be really competitive in this diclipline. The way for Mali suceed is to strictly exploit their boni, which means amassing faith and gold. The latter is their perfect replacement for production - they can easily produce tons of gold and the purchses they make with it are cheaper on top. And focussing on their boni means also letting other settle the fertile spots, because those big cities give the Mali perfet room for sending their carawans. Leaving good spots to others surely bites with an aggressive approach of playing, but that's not the Mali style in civ anyway (plus they have the advantage that settling the desert can happen without interfering anyone else and is less likely to enrage neighbours)

 

[Loderingo]

Surely if the plan is to get as many cities as possible, then the best strategy is to settle the fertile land first then take the deserts which no-one else wants later

 

[ezzlar]

Mali shouldn´t settle in deep desert but on the fringes. They can then use their UA while working "normal" tiles. But what I can see the city placement advisor doesn´t seem to count the UA?

 

[Troy Bruckner]

Mali shouldn´t settle in deep desert but on the fringes. They can then use their UA while working "normal" tiles. But what I can see the city placement advisor doesn´t seem to count the UA?

A deep desert with mostly flat desert tiles is key to Mali success.

 

[Infixo]

A desert is coded as a type of terrain and the settling algorithm in the game doesn't support a terrain, i.e. one cannot differentiate a decision based on what terrain is around. It supports features, resources, yields, etc. but not terrains.
Same problem is with Pachacuti - he cannot check for mountains.

 

[Pfeffersack]

After experiencing another time Mali AI not getting really into a game I remain convinced that settling desert first is the better approach for them - at least in the hands of the AI. A skilled human - who is able to beeline for the best places anyway, but can also avoid problems like barbs on the way- might suceed with stealing others best settling locations first, then settling desert afterwards. But the AI is doomed this way - it might not even be so much the wasting of their boni, it seems to be mainly a problem of loosing simply speed and security. Mali starting in the deep desert often means that the "generic good" settling spots are away some tiles and reaching them costs both time and increases the risk of getting cought by barbs (which are quite likely to show up in deserted desert to make that pun). In the example below that Mali could have easily secured 2-3 iron sources by settling desert.