The data has 5632 games, we used pr30+ games as a compromise between high level and number of games. 35+ would be ideal but would leave us with a data set too small to draw any conclusion;
also, mirrors were excluded from the data (ask @Buckethead why).
The following table includes all rated 1v1 games played in EP in the last 6 months, where both players are above pr30
- The popularity row represents the percentage of games a given civ was played by one of the players;
- Most played MU was France vs Russia, with 240 games, so is safe to assume that 43,3% winrate close to true;
- Least played MU was Sioux vs Aztec, with only 5 games, so that 40% winrate has a very low confidence level and shouldnt really be a parameter for anything;
For easier readability, each civ win rate in decrescent order:
Sadly the dataset is too small to give any confidence in specific Match ups. When we look at spain vs aztec we see a win rate of 85,7%, but looking further into the data we see that only 7 games where played in this mu, with a score of 6-1 in favor of spain. Hard to take any solid conclusions out of this. The raw data is bellow, if someone is interested in looking into the details of anything.