The game will have a potential matchup balance problem due to being more asymmetric that AOE2 and this problem will increase when the number of civilizations increases (source)
Initially the game will have 8 factions which means 28 matchups (excluding mirrors).
When the game reaches 12 factions it will have 66 matchups (excluding mirrors).
Even if a civ is balanced overall with a close to 50% winrate it might be OP in one matchup and UP in another. Say that Chinese overall winrate is 50% (balanced) but Chinese vs Briton winrate is 41% while Chinese vs Mongol winrate is 59%.
Trying to balance this in the normal way would be impossible. Any change you make to improve one matchup could make several other matchups worse.
So what is the solution?
The easiest solution is to have an inbuilt automatic balance correction that is matchup specific. The disadvantaged civilization gets extra starting resources but only in the matchup where it is weak.
The disadvantaged civ get 100 extra starting food and gold if the matchup specific winrate is less than 45%. If the winrate is less than 40% it gets 150 extra wood/gold etc.
This way the developers can focus on adjusting civilizations that are too strong or too weak overall, while the game itself corrects balance problems that only occurs in specific matchups.
The civs are mainly different in their landmarks and eco bonuses.
Military wise it’s basically the same except maybe 1 outliner, in which I’d say Age2 was more imbalanced in that regard (turks no spears big sad).
The game is waaay more symmetrical than you would like it to be, which makes balance mistakes really really unlikely.
The only outliner we have seen so far is mongols in specific, and they seem to be really underwhelming from what we could see (relying on gimmicks to work and really unstable).
Giving a civ artificially different starting ressources depending on matchup stats is an absolutely horrible idea.
It makes the game more inconsistent and punishes any temporal stats outliner cause of meta developement, making it harder to figure the game out in the grand sceme of things.
It doesn’t take into consideration the skill level required to play each civilization. As an example the English are considered a beginner civilization so it’s possible that, on low to mid range ELO, they will have a very high win rate compared to a more involved civilization. This doesn’t mean it’s a better civilization and it certainly doesn’t mean that everyone needs to be handicapped for using it, especially on higher ELO where they might actually be the worst.
I did some stat about balance on aoe3 and you can’t use the overall winrate to determine which civ is the strongest because it depend of the level of each player (somes civ perform well at low level), also their are some biasis like, “if i am better to my opponent, i will play a weaker civ, then i will inflate the winrate of this civ” (it’s something we have found in aoe3 because somes of the top players had a better winrate with low/“not main” civ because they played mostly against lower player with them).
I see where you are coming from, but the civs differ a lot more on the way they set up their economies and they all seem to have solid boom potential which is nice.
Some match ups may slightly favour one civ over another but its not like they will be highly one sided as civs have the same core mechanics and all of them have access to the same core units as well.
For this solution to work you should not look at the winrate of every player. Instead just calculate the winrate each time a top 1% ELO player plays another top 1% ELO player. Then let the game update which civ that gets a matchup balance bonus once per week or once per month.
This solution becomes especially important if future civilizations get unit types that are not currently in the game. Which I do hope since it will be boring if everything is just archer/knight/pikeman/swordsman.
The civilizations needs to have larger difference in units in order to keep the game interesting since just having different economic bonuses gets stale fast.