Tweet
My personal AI tweaking
CTP2 was my first turn based game, and my first civ builder. Ive been playing the game since it came out, winning on a regular basis at all levels.
Now, ive tried some of the tweaks suggested by other users, and never really saw much difference. Then I started thinking, with all the various tweaks, which ones actually help, and do some possibly cancel out others. So I picked out a couple that seemed likely to make a difference and used them. For instance:
*I made the following changes in strategies.txt
OffensiveGarrisonCount 8
DefensiveGarrisonCount 2
RangedGarrisonCount 2
*I changed the above in all instances in strategies.txt
*Also, I upped the maxexec and maxeval settings for seige(which apparently affects attempts to take a city)
My reasoning behind both of the above was to get the AI to defend its own cities well, while also considering taking mine(if we arent at peace)
I chose the numbers at random, no real thought went into it. 8-2-2 seemed neat and tidy.
I then started a new game and did notice the AI beefed up its defense of cities, but that was about all.
However, when it had many cities, as the Phoenicians did in my game, most of the map that wasnt controlled by me, many of the cities were poorly defended.
I attributed this to the mere fact that the AI couldnt afford to put the full complement of units in all the cities it had conquered. So, I thought, what if the AI could afford to churn out enough units to field a sizeable defensive force and also an offensive one.
This is where I tested my theory. I edited the DiffDB.txt file to give the AI a ton of gold and PW. I edited the entries for "very hard" as thats where I keep the difficulty setting.
I started a new game on an island map with several large islands. The Japanese had one city on my island and all of another one. I waited until I could field 2 stacks of 12 hoplites each and then attacked their city on my island to provoke them. The city was defended by 12 fortified units, and my attack failed. I sent in my second 12 stack a few turns later and they had Samurai, my hoplites were slaughtered. Now, suddenly, here comes some 4 and 8 stacks toward my cities. Since I used all my units to make the 2 stacks, I had virtually no defense, so the Japanese took 2 cities right away. I played for about another 10 minutes, and noticed the AI was rush buying wonders and its cities were growing fast. I will continue tonite to see if the AI can launch an offensive.
Now, I know loading up the AI with gold and PW is not the solution we desire, but maybe a proper tweaking of settings to allow the AI to grow and prosper much better would help.
Of course, this is just one way to approach the AI shortcomings. If enough of us test various theories, we may eventually come up with something.
Now, ive tried some of the tweaks suggested by other users, and never really saw much difference. Then I started thinking, with all the various tweaks, which ones actually help, and do some possibly cancel out others. So I picked out a couple that seemed likely to make a difference and used them. For instance:
*I made the following changes in strategies.txt
OffensiveGarrisonCount 8
DefensiveGarrisonCount 2
RangedGarrisonCount 2
*I changed the above in all instances in strategies.txt
*Also, I upped the maxexec and maxeval settings for seige(which apparently affects attempts to take a city)
My reasoning behind both of the above was to get the AI to defend its own cities well, while also considering taking mine(if we arent at peace)
I chose the numbers at random, no real thought went into it. 8-2-2 seemed neat and tidy.
I then started a new game and did notice the AI beefed up its defense of cities, but that was about all.
However, when it had many cities, as the Phoenicians did in my game, most of the map that wasnt controlled by me, many of the cities were poorly defended.
I attributed this to the mere fact that the AI couldnt afford to put the full complement of units in all the cities it had conquered. So, I thought, what if the AI could afford to churn out enough units to field a sizeable defensive force and also an offensive one.
This is where I tested my theory. I edited the DiffDB.txt file to give the AI a ton of gold and PW. I edited the entries for "very hard" as thats where I keep the difficulty setting.
I started a new game on an island map with several large islands. The Japanese had one city on my island and all of another one. I waited until I could field 2 stacks of 12 hoplites each and then attacked their city on my island to provoke them. The city was defended by 12 fortified units, and my attack failed. I sent in my second 12 stack a few turns later and they had Samurai, my hoplites were slaughtered. Now, suddenly, here comes some 4 and 8 stacks toward my cities. Since I used all my units to make the 2 stacks, I had virtually no defense, so the Japanese took 2 cities right away. I played for about another 10 minutes, and noticed the AI was rush buying wonders and its cities were growing fast. I will continue tonite to see if the AI can launch an offensive.
Now, I know loading up the AI with gold and PW is not the solution we desire, but maybe a proper tweaking of settings to allow the AI to grow and prosper much better would help.
Of course, this is just one way to approach the AI shortcomings. If enough of us test various theories, we may eventually come up with something.
Comment