@Everyone who says "the higher the playrate the lower the winrate"

Why is that? This is one of the most popular "balance myths" on the boards and I think that while I can see where people are coming from it's still not that simple.

The argument itself is rather straightforward, when a champion has a high playrate it's rather likely that it's a FOTM (or: meta) champion and therefore attracts many "new" (to the champion) players. Inexperienced perform worse than moderately experienced players therefore those popular champions have a lower winrate than what would accurately represent their strength. Right?

Wrong - at least sometimes. There are certain champions that attract a very dedicated playerbase eg Gangplank and Riven - they have a more experienced playerbase and should therefore have a higher winrate than their actual strength is. If you were to look at Riven's playrate you'd think that most Riven players are inexperienced - but that's just factually wrong, in fact Riven has the 4th most experienced playerbase of all champions, despite having a 16% pickrate.

So what is the solution - just looking at "avg. games played"? Probably not. While it's of course better than the playrate itself, you still got to remember that those with 1000 games on a champion should have a winrate only slightly above 50% winrate as they should have reached their true elo (assuming that they mostly play that champion). That's why I suggest looking at the learning curve of champions, ideally on leagueofgraphs.com as they have the most reliable data (champion.gg's samplesize is notoriously inaccurate and op.gg is generally good but recently lost the "trends" section which is super lame).

Because a quick look at the learning curve tells you a lot: you can compare how beginners perform to how they should perform (Vayne beginners have a 50% winrate? Sounds OP!) or how experienced players perform compared how they should perform (Vayne mains barely manage to get above 50%? Sounds OP).

Data analysis is a very complex topic so the learning curve is still influenced by other factors (does a Yasuo beginner have as much flex/normals as a Maokai beginner on that champion? Does a higher playrate also mean that people are better at playing against that champion? Does high playrate mean that the more people play the champion correctly as there is a popular build or similiar things? Is the champion very popular as an offrole champion so that the avg. games played is very high despite the players being inexperienced with the role?) but overall we have little accessible data, so let's try to use the most accurate that we can get our hands on.

tl;dr: high playrate doesn't equal inexperienced players, use learning curve on leagueofgraphs.com instead.