[2023] How this is even possible?

Using an extreme example (with easy numbers):

3 Players play U-190.

Player A plays Morning and Evenings (10 battles each)


Player B plays Mornings only (10 battles)


Player C plays Evenings only (10 battles)



All Morning Matches are Player A vs B


All Evening Matches are Player A vs C

Player A loses all their matches. Player B and C win all their matches.

End result:

Player A - 0% WinRate (20 battles)


Player B - 100% WinRate (10 battles)


Player C - 100% WinRate (10 battles)

Overall WR by Battles for the U-190 with the above information is 50%...

...but wowsnumbers is displaying WinRate by account, meaning the 3 accounts to pull data from would show a 66% overall WinRate.

-

There is such a mass of profile data on the site that the numbers do fall close to norms/expectations, but as mentioned above there are accounts with Hidden Stats which won't be tracked and there can be a myriad of other minor variances in such a large-scale project.