A quick question on your most recent test -- whose results are indeed disturbing (I'm an ex-programmer and can't even imagine why anyone would design an algorithm in such a fashion) -- how were the units flagged re: O/D?
The same happened in the earlier tests. I changed the Longbowman to Skirmisher, and the AI seemed to prefer it over the Archer, even if the Archer had higher stats. In every test that featured both units, the AI prefered building the weaker Skirmisher unit over the Archer unit, even if both units cost the same.
I then believed it had something to do with the fact that the Longbowman is superior to the Archer in the regular game, and that AI-preferences might be hardcoded to follow the epic-game unit upgrade paths.
However, in the last test with the added and (thus "clean") Spearman 2 and Pikeman 2 units, the AI still prefered the Pikeman 2 to the Spearman 2, regarless of it's strenght.
In the both tests, on first turn, when all cities were undefended, it would build the Spearman 2 unit. Even when the Pikeman 2 unit had 1-3-1 stats, and costed the same. That's stupid enough as it is. However, after those first units, it would almost solely build the stronger pikeman 2 units.
But in the second test, with the Spearman 2 units being the strongest, it still followed the same behaviour. First a Spearman 2, then a bunch of Pikemen 2. Eventhough the Pikemen 2 was weaker. It still prefered it.
There must be something hardcoded in there somewhere. But I can't figure out what it is. Both units were "clean". Why did it still prefer one over the other, even if it's weaker?
Comment