Predictability: good or bad?

[DrSpike]

Originally posted by LDiCesare
Anyway, I think I agree with you on the fact that someone has been lazy. The who we can't know for sure, but that probably doesn't matter much.

Agreed.

Either way it doesn't seem likely to me that the reason was the impact on size, but laziness somewhere.

[rah]

I don't see it as lazy, (for huts and specials), I think that for civ I, no one thougth the need for it be unpredictable and figured a pattern would be neat. CivII did it the same way probably because no one thought to change it.

If you think about it, I can't remember anyone complaining about it prior to MP coming out since in SP its to the Human's advantage because we understand the patterns and know how to take advantage of them. It's obvious that the AI doesn't.

[Six Thousand Year Old Man]

Originally posted by rah
It's obvious that the AI doesn't.

One thing we haven't figured out is how the AI thinks. I'm not even talking about combat, here... things like, why build on this grassland square and not that one? Is it random? Does the AI count the arrows/sheaves/shields that a site would generate, and build on the best one within x squares of its settler? I postulated in one of the hut threads that a nomad/tribe result might depend on the 'quality' of the city radius, as determined by the program, and not just the terrain of the site itself. But then, I got confused and gave up that line of thinking

Anyway... some of the 'predictable' things, I use. Obviously I look for good sites based on visible specials. I seldom mine hidden specials. I oedo to monarchy, but usually forget to use it later on. I black click sometimes. I guess that I'd prefer a more random game for 2 reasons: It might make the game more exciting (imagine finding a OCC site with 7 specials, say, or getting stuck on a continent with no specials, while the AI has one loaded with 'em), and, in MP, I suspect the predictability would give those who know the game better than I do, even more of an advantage.

[rah]

Originally posted by Six Thousand Year Old Man
One thing we haven't figured out is how the AI thinks.

After seeing most AI cities placed one square out of reach of the special, I'd have to say that however it does determine it, isn't very good. And actually quite terrible. It probably just settles the first grassland it finds. It's fun to watch an ai settler wander over rough terrain.

[LDiCesare]

rah's pattern explanation is mine too: Nobody thought it worth doing things different from civ1.
As for settlers choices, I think the terrain is chosen, but not its actual value - that is if you mod terrain so that grassland gives no food but mountains do, the ai will still not settle on mountains. I think the program gives some silly advice at the lowest levels about "this looks like a nice place to settle". Maybe playing chieftain might give insights about the ai's settling patterns?