win rates can be very different at different levels of play and even when they happen to be at the same side (which is often not the case some civs have low win rates at low level and low winrates at high levels) e.g both positive win rates it is not necessarily for the same reasons.
The screenshot above is from aoe4 world, but what Heftydogg did not tell you is that the sample size is small conqueror level (for which he checked win rates) even the website itself tells you " low sample size it is not a good idea to post on reddit" I guess they should add “nor on forums”. On top of that there are lot of variables that bias the results including selection bias e.g some civs are picked more on maps in which they tend to perform well, if players were given random civs to play and were equally good at all civs then win rates would be a much better metric.
Also there is an issue I noticed from checking “the streamed matches data” and while stremed matches are only a subsample of the total high level games is that most of the time the winner is the guy with a bit higher ELO regardless of the matchup and who is favored so if you get more " higher Elo English vs lower Elo Ottomans" games than " Higher ELO Otto vs Lower Elo English" games you will get a misleading winrate.
It is hard to infer causality without very in depth scrutiny.
Also notice that the current win rates for ranked 1v1 are these (the screenshot of Hefydogg is relatively old). In his screenshot you see 54.3% for conqueror (151 games). Right now I see 52.7% with 262 games. And remember this is an estimate so it should be read as 52.7% +/- x with some degree of confidence but the website does not show x (you could calculate it). (see the fresh win rates below)
Then you have the problem conqueror does not represent the “high level of plays”. So at the end of the day we hardly have any samples for high level plays. I also add the fact that those numbers are estimates so if you read.
Lastly you could have people that are not main ottomans playing ottomans and since the sample is small they will have a big impact on the stats.
This is the win rate table for Conqueror games.
And these are the current win rates for different ranks
(I don’t know why the sum of number of games across bronze, silver don’t sum to the amount for the catergory “All”. Anyway I just took the numbers from the website.
Now imagine Conqueror Malians players come to forum (which is way more valid than Gold players doing the same thing) to complain about Abbasids being OP and needing big nerfs because Malians have 46% win rate against them in conqueror. Meanwhile Abbasids have low overall win rate.
Then you get things like Malians having a 44% win rate against English in Gold but 60% in the current Conqueror sample. They also have a 36% win rate against Ottomans in Gold but 49.5% in the current Conqueror and so on. So should Malian players from Gold come complain about English being unbeatable ?
The number 1 way to get better ranks in the game (and by a large margin) is to have a better macro even with 0 strategy and after turning off the brain and not thinking at all and the threads made about imbalance in Gold/Plat/Diamond whatever (I would say even in low conqueror btw) are mostly pointless