Txurce
Deity
Are you playing, or running it on autopilot with one of the editors?
[151.9] Giant AI Bug
- Thread starter Seek
- Start date
Seek
Deity
- Joined
- Aug 7, 2010
- Messages
- 3,773
110 turns on 147.4 - no bug. I'm strongly suspecting the bug arose around 148-9, and has not in fact been around for months. Which is good news, I think!
Txurce, have you tried Thal's suggestion to use the Tuner to switch players in your current games?
Txurce, have you tried Thal's suggestion to use the Tuner to switch players in your current games?
Txurce
Deity
v148 is when we had the first reports of certain AI basically going on strike.
I wasn't sure what I do once that happens. Could you elaborate a bit? I'm in a hotseat game today, but would try it tomorrow.
Thalassicus
Bytes and Nibblers
This is the likely cause, a change made primarily to the yield library. I don't know how to solve it while maintaining the player:method format. I might have to revert everything to the old format... which will be a pain, because I've made more changes since then, so it will take hours to revert everything. The find-replace terms used to go one way are much simpler than the ones needed to go back.
Still... this is important enough that it's worth a try.
Thalassicus
Bytes and Nibblers
Actually... now that I think about it, I don't believe that's the issue. I was only able to replace the player-based functions, not the city-based ones. The problem you describe is with individual cities; we're back to square one.
Txurce
Deity
Commenting peripherally: in my game where Istanbul was size 2 on t62, by t100 it was size 11, despite having added 2 other cities (without building the HG, either). There was a restart at t35 (no effect), and one at t62.
Civs in general are expanding more slowly, but in all fairness it's more in line with the human player, which is actually what we were trying to achieve at one time.
Civs in general are expanding more slowly, but in all fairness it's more in line with the human player, which is actually what we were trying to achieve at one time.
Seek
Deity
- Joined
- Aug 7, 2010
- Messages
- 3,773
Just check an AI's cities and see if the yields displayed add up to what they should be from the hover tooltip (ie, 4 from terrain, 2 from buildings, etc.)
Sorry for the late response, but I missed this post earlier, somehow.
rfxmills
Warlord
If the AI performance really drops off a cliff would it be easier to see in infoaddict?
Txurce
Deity
IA would probably tell you there's a problem, like low city count, population, production, etc. Going in with FT allows you to see exactly what is causing the problem.
rfxmills
Warlord
If you see a problem in IA, switch user and the yields jump back up, you will know that the yield library might have reset?
Txurce
Deity
Now that's troubleshooting! You mean using both IA and FT, right?
Thalassicus
Bytes and Nibblers
I use git, but I don't think it can help here. The format changed, then I wrote new code using the new format. Reverting to an old version would just delete the new code. It doesn't matter either way because the format of the "city yield" code is the same as before. Only the Game class and Player class functions changed format.
Thalassicus
Bytes and Nibblers
Does the problem only happen in AI capital cities, or others as well? It might be a mistake with the AI capital bonus yields. I changed how they work in this version:
rfxmills
Warlord
Haha yeah, that year as a QA tester wasn't as much a waste of time as I thought ;-)
Txurce
Deity
It would seem to be the capital, although that's partly because it's where you're most likely to notice it. I'll try to keep an eye on it.
Is there a reason why they would respond differently on occasion after restarts? (The Ottomans going from a size-two capital on t62 to 7 cities (#2) in t130 is mindboggling.)
Similar threads
