@Riot, How Do You Measure Game Balance?

I don't think winrate is a good measure of balance* at all. To illustrate why, consider the following thought experiment:

Let's say there's a champion with a passive that gives 500g to every member of his team or the enemy team (50/50 odds on which). It really doesn't matter what the rest of his abilities are (assume a 3 hit proc, %maxHP damage, some kind of dynamic mobility, and an ult that bends the rules as we know them) because whichever team gets the bonus gold at the start of the game will almost surely win with such an advantage. So his winrate will be around 50%. But he's not really balanced because half the time his team is steamrolling and the other half they're being steamrolled. If you're really addicted to winrates, you could argue that he's a great champion because he perfectly balances every other champion in the same match! Hooray!

This is a flatly ridiculous example, but there are a lot of champions which people hate because they feel like whoever is playing that champion has an unfair share of control over the outcome of the match. Sometimes these champions even sit below the 50% winrate mark because they can be very hard to play. If I'm playing against one, however, I don't want the outcome to be solely in the hands of an RNG (as in the absurd example above) or largely in the hands of the other player (even if their champion has a really high skillcap)--I want a fair share of it to be in my hands.

Winrates can be a particularly deceptive red herring because they don't measure the total impact of a champion on a game, only whether or not that champion's presence tends to lead to more victories than defeats. So, Riot, how do you measure champion impact? And how does that contribute to champion balance?

*Some people may think "balance" is defined as winrate, in which case, substitute "balance" for "balanced for healthy gameplay" in the post and continue as normal.