Just out of curiosity, what formula was use before?iYieldHandicap = (iHandicap * iEra * iEra);
Just out of curiosity, what formula was use before?iYieldHandicap = (iHandicap * iEra * iEra);
Three differences:Just out of curiosity, what formula was use before?
Did you change the deity handicap from 7 to 5 or is it still 7 ?
Would it be hard to release some alternate .dlls with formulas we can test?Starts at 5, scales down from there. We’ll see. I had some insane games with atomic around 1400 on Deity at 7 so I dropped it. It may need a higher base value though. Dunno. Thankfully if it feels too easy you can raise it back up in DifficultyMod.xml
G
Precisely what I was thinking. Around Renaissance, AI will be as strong as usual, but it needs to recover from the worse beginning.Would it be hard to release some alternate .dlls with formulas we can test?
Also remember to play out your games guys. If the AIs are supposed to be slower pre-ren and faster post-ren. Obviously there's a balance, but if you don't play it out we can't actually know if it's working on the other end. (It's conceivable that the balance can completely switch, where the AI falls 5-10 techs behind and then speeds past you to the finish line.)
I play life on Quick Speed.
Really funny.
G
There was a typo in our computations, so current difficulty level is absurdly low. I would say that the AI lost 2 difficulty level in this version. Except for Tall AIs.
Would it be hard to release some alternate .dlls with formulas we can test?
Also remember to play out your games guys. If the AIs are supposed to be slower pre-ren and faster post-ren. Obviously there's a balance, but if you don't play it out we can't actually know if it's working on the other end. (It's conceivable that the balance can completely switch, where the AI falls 5-10 techs behind and then speeds past you to the finish line.)
The problem I had anticipated with this change is that the easiest way to win for players is to get an early advantage and use it to plow through AIs; if the AI systematically falls behind in tech early game it's a death sentence against players who are willing to use the opportunity.
I think Moi Magnus offered one very flexible formula ax^2 + bx, written as a * iEra * (iEra + b). It gives a steepness intermediate between a linear and a quadratic curve.I agree that the danger is for players to snowball early, and be in position to hang on against a late-blooming AI. I'd add that an AI that picks up too much speed at the end isn't any fun, either. But I think it's still worth experimenting with the slider (so to speak) until either a sweet spot is found, or the effort is dropped as a worthy but failed experiment.
Am I the only one that can't build plantations on tea? Bad install or something?
Reinstalling and trying again.
The ideal is for the AI to be a reasonable challenge for the player at every stage of the game, but failing that, I think it’s better to lean toward the AI being strong early and weak late rather than the other way around. Starting behind the AI and catching up leads to come-from-behind victories, which can be exciting. Whereas effortlessly surpassing the AI in the early game only for them to catch up and overtake you out of nowhere later on is less fun.
I disagree. If we can't reach the ideal of challenging at every stage, I'd rather weaker AI early, and stronger late, so we don't have the all too common case where it's no point in finishing the game after modern era bc it's clear that we're going to win.
Weaker (if it must be the case) earlier still means that the AI you don't wipe out early will stick around, evaluate the situation the best they can and come back strong as the game progresses.
I also agree that a weaker early AI game is more interesting to me then a strong early AI game. Also won't wiping out AI early cause the survivors to band together to take you down later?