Snipped from:
AI
-----------------------------
Soren was in charge of the AI and it took "7 or 8 months of his time to write it about half writing and then half testing." He'd set the AI to play against itself for ten minutes, and then rewrite as necc. He also tried playing against it himself, of course.
----
>>>
With this methodology, I can see how and why the AI is what it is. With only 4 months to do coding, I believe it is a minor miracle that the AI is as good as it is and a testament to the brilliance of the programmer. (Good in the sense that it can beat 20% of players on a level playing field.) Sure the AI does stupid things, but with one person coding and one person testing, many loopholes will never be found. Also there is the phenomenon that fixing one stupid behavior often leads to another, even more stupid behavior.
I am sure there was a lot more nipping and tucking after the primary engine was written, but the die was cast. It seems like it is too bad that an experienced gamer was not in the early AI development loop for testing and for ideas. Mr. Johnson may be a brilliant programmer, but I doubt he is in the top 20% of Civ III players or any 4x type game. (More likely the top 40% or top 50%.) Having a top flight Civ II player involved early, probably would have been a tremendous resource for Mr. Johnson.
My compliments to Mr. Johnson for coding the Civ III AI with what I see as limited time and resources.
Cheers.
AI
-----------------------------
Soren was in charge of the AI and it took "7 or 8 months of his time to write it about half writing and then half testing." He'd set the AI to play against itself for ten minutes, and then rewrite as necc. He also tried playing against it himself, of course.
----
>>>
With this methodology, I can see how and why the AI is what it is. With only 4 months to do coding, I believe it is a minor miracle that the AI is as good as it is and a testament to the brilliance of the programmer. (Good in the sense that it can beat 20% of players on a level playing field.) Sure the AI does stupid things, but with one person coding and one person testing, many loopholes will never be found. Also there is the phenomenon that fixing one stupid behavior often leads to another, even more stupid behavior.
I am sure there was a lot more nipping and tucking after the primary engine was written, but the die was cast. It seems like it is too bad that an experienced gamer was not in the early AI development loop for testing and for ideas. Mr. Johnson may be a brilliant programmer, but I doubt he is in the top 20% of Civ III players or any 4x type game. (More likely the top 40% or top 50%.) Having a top flight Civ II player involved early, probably would have been a tremendous resource for Mr. Johnson.
My compliments to Mr. Johnson for coding the Civ III AI with what I see as limited time and resources.
Cheers.
Comment