AI weaker in non standard games?

[tobyr]

I doubt that "the civ engine has been especially designed for the normal game"; rather, the civ engine has primarily been debugged,. tested and tried out using the normal games. It surely takes a lot of time to tune the AI, to make sure it is working well. It is not cost-effective to also test it with every variation in land size, world-creation, rules options, let alone "typical" changes in rules.txt. So inevitably it works best in the standard game.

Please note that in the ToT fantasy game, it is doubtful that the AI plays as well as in the standard game.

- toby


------------------
toby robison
criticalpaths@mindspring.com

[SandMonkey]

The aliens play a good game in the extended game, usually. And heck, you can win as the alienbs in extended game (I've done it), and the AI was fairly close in tech. I think, like toby said, that the AI wasn't designed to play on worlds like this, so they might play a little under par (which is fine by me, i still haven't gotten the hang of those worlds!)

------------------
SandMonkey

"Shut up brain or I'll stab you with a qtip"
-Homer Simpson

"Ecky ecky ecky!"
"It's just a flesh wound!"
- Monty Python and the Holy Grail

Check out my 1602 A.D. site

[Spirit]

Ofcourse if the AI would only be playing a little under par, it would be normal. Yet the differences are according to my own experience a little bit too big.
Surely the rate at which the computer expands his empire shouldn't be influenced by the rules.txt as long as the settler-type units are the same and yet this is my greatest concern. There's quite a big difference in cities, when comparing my games. (see my original post, )

[Spirit]

Recently I began to wonder if the AI really was playing worse in non standard games e.g.(Games with a radically different rules.txt file.)

Check this out:

In the Extnd. Orgininal game (ToT) around 1850 I had N+S amerika and 2 AI players had taken 85% of europe,azia and afrika (conquering 2 others). Research was about equal.

Comparing this to the Lalande game (ToT)
again 1850, but now I have half the "normal" world (Funestis), half orbit (not so much) and half Naumachia, for a total of 160 cities,
the 5 computer players share a common interest of about 50 cities and are only half way the tech-chain, (I'm building my spaceship). I didn't have a great start, just expanded normally, no major wars, etc.

So, as I have had equal experiences with other scenarios, I wonder whether or not the civ engine has been especially designed for the normal game, if it really originated from civ 1, I'm not surprised,(this isn't the case, is it?) , yet if it didn't, why is it playing so poorly?