While I partly agree on team games (3vs3+), this assumption doesn't work out in 1vs1 games as all the ladders are separated, hence the ladder with the most players should've the highest winning rate, as we have most players competing there. Yet this is not the case and balance problems in 1vs1 amplify in team games, that's why 2vs2 is in even a worse state.
In 1vs1 you just wait longer to get a proper ranked opponent as allies, it doesn't mean you get more noobs to bash - rather less.
What it is interesting is that while both 3v3 and 4v4 were shaken by the introduction of UKF, turning the tides in favour of allies, 4v4 has gone back to be slightly Axis favoured while 3v3 remains allied favoured.
Also, we need AT rankings rather than randoms for teamgames.
The only way to fix this is to simply include the majority of the community and get a overall faction winrate but this won't represent high level play but will give an accurate idea of how the factions stack up in the game mode or to filter games that are outside the arbitrary high skill bracket.
There's a point on which you have to draw a line, were it's not a balance problem more than a lack of knowledge or ability of a player.
I'll say that using W/L as a gauge of balance is stupid, unless results are heavily skewed to one side or the other (see Axis W/L for 3v3+ and still 2 huge reason were playerbase and MM)
Extending to top500 or even top2000 won't do shit, unless you can filter matches based on similar ELO or rank difference of not more than 100.
Relic Entertainment
보드카 중대 
