Winrates DO mean something.

But they are NOT a straight indicator of whether a champion is broken or not.

They do, however, carry valuable information.

Winrate and Pickrate - Inverse Relationship

To be brief: Take Heimerdinger, who seems to always be well above 50% winrate. Why is this? Does it mean he's broken? But so few people play him! That's the point. So few play him that the few who do are generally very experienced heimerdinger mains.

This is actually an observable trend. Regardless of a champion's balance state, watch their popularity; Usually, as a champion becomes more popular, their winrate drops. And the contrary is also true; usually, a very under-played champion has an above average winrate due simply to being played only by people who know the champion very well. (Mains.)

There are exceptions. When a champion is so bad and so underplayed that even their mains can't bring their winrate much higher (think something like Urgot, maybe.) Or when a champion is so blatantly broken that it's easy to stomp with him/her, even as a player who's just picked them up. (Think Ahri after her recent changes that shot her to 56-58% winrate, before the subsequent nerfs.)

That's all. Winrates can give us valuable information, so we don't need to discard them as useless.
But saying a broken champion is balanced because their winrate is poor (while they're also incredibly popular) is an invalid argument. A champion can be overwhelmingly broken, but still suffer a low winrate - because so many people pick that champion without knowing how to play them, simply because they've heard they're OP.

Thanks.

Counterpicking Influence

[{quoted}](name=Malix Farwin,realm=NA,application-id=3ErqAdtq,discussion-id=TYnpEVZd,comment-id=0003,timestamp=2015-03-11T12:14:45.876+0000)

One key thing you did not mention about Win rates. They are also based on how well the champions are doing against the top meta picks. For example, annie's win rate is so high because she counters a ton of the current meta picks. However, if the meta were to shift more around champs like galio for example her win rate would drop.

^ Great point.

Implications of winrate & pickrate on champion skillfloors & skillcaps

[{quoted}](name=The AFK Farmer,realm=NA,application-id=3ErqAdtq,discussion-id=TYnpEVZd,comment-id=0002,timestamp=2015-03-11T12:05:13.694+0000)

Surprised no one corrected you. I will then. Statistics from op.gg.

All Leagues, Last Month:

: 33% pick rate; 53% win rate. : 30% pick rate; 51% win rate. : 30% pick rate; 47% win rate. : 26% pick rate; 52% win rate. : 24% pick rate; 50% win rate 23% pick rate; 46% win rate : 21.5% pick rate; 52% win rate : 20% pick rate; 53% win rate : 18% pick rate; 50% win rate

As you can see, only Thresh and Lee Sin have sub 50% win rates. That could be because poor players are picking them, or it could just be the nature of their kits. Even in higher tiers, though, they maintain sub 50% win rates. In Platinum Lee Sin still has a 47% win rate and Thresh has a 48%. In Diamond, they both have 49% win rates.

I'm not seeing a high popularity-low winrate trend.

The end of the post addressed instances where champions are both high pickrate and high winrate (or maintain a clean 50% winrate despite being incredibly popular.) This is usually an indicator that the champion is actually overpowered. Not having a high winrate alone. But maintaining one while being a very popular pick.
The champions that are broken, then, become even likelier to be picked, and so their pick rate increases.

Basically what I'm saying is that, if you're going by the most popular recent champions, then you're liable to find candidates for somewhat overbearing champions. For example. I really like Jinx's kit, and her strength despite being immobile. But it's possible she's become slightly too good lately due to the continuous chain of nerfs to whatever other ADCs rose above her. Or, to be more accurate, she might have become slightly too reliable.

Other champions on that list kind of speak for themselves. Riven, for example. I feel that despite having an incredibly high skillcap, she's easy to play in a binary feast-or-famine matter. All-in, press every button every time they show up to lane; it's very likely that either you die and keep feeding, or you get the first kill and keep killing them repeatedly. Her skillful players (like BoxBox) can minimize the damage from all-ins that go wrong, or get kills where other Riven players wouldn't have been able to. But then there will also be Riven players who just consistently feed. Binary, feast or famine, low skill floor, yet high skill cap; balancing at 50% winrate.

Notice the implications on skillcap given by the annie vs lee statistics, by the way. One presses all of their buttons and hopes you die. The other, Lee, has a very high mechanical skillcap. Incidentally, I think Lee is a more overbearing, overloaded champion overall, but the point remains - they're both very strong, but the one with the high skillfloor (Lee) suffers in winrate when tons of people play him, where as the one who is relatively mechanically simple (Annie) maintains a high winrate despite her popularity, due to how easy she is to use. Although I do think she has a high skill cap as well. But incredibly low skillfloor.

Perhaps the simplest example is how close in popularity Thresh and Leona are, and yet his popularity is a bit higher, and his winrate is a bit lower. Why? Both bring lots of hard CC and tankiness to a team. Both have similar engages. Overall, Thresh seems to offer more utility, but have a higher skillcap. The one that is more mechanically difficult is, in my opinion Thresh; he therefore suffers a lower winrate by default due to the massive amount of people picking him and not playing him to his full potential. Whereas many of those players might have fared better with a more straightforward kit like Leona's.