Title. All the time here in gameplay I see complaints like "X champion is only 49% winrate pls buff!" or "X champion is 51% winrate! NERF NAO!!1" Winrate cannot be exactly 50%, be happy with 49% or 51%. Jeez they don't need buffed for one freaking percent. 40% winrate? Sure that's kind of bad. 60%? Dang! But when you're complaining about 55% or 45% I just want to report your post for spam.
Randomonium2 votesWin rate is something that a lot of people don't understand because win rate is strongly influenced by champion popularity and perceived strength.
A champion may have below a 50% win rate while still being balance because they are a very popular champion. Popular champions typical have a lot of people play them who aren't very experienced with the champion which typically drives the win rate of the champion down.
A better indication of champion health is to look at win rate among players who have 100+ games with a champion. This can give a better indication of what a champion's "upper ceiling" is and give some good insight into how balanced a champion really is. The problem with just looking at these players though is the sample size is much smaller so the chance for inaccuracy is much higher.
Really to get a true sense of balance you need to look at all the stats and how those stats change with player experience. This is why I love sites like champion.gg.
TL;DR you're right but I don't think you went deep enough into the "why" aspect of it.
Randomonium2 votesYes, yes really. People who are less experienced with a champion will generally have a lower win rate than people who "main" the champion. If the champion has a high popularity the people who "main" the champion make up a smaller percentage of the total games played with that champion. This drives the overall win rate of the champion down.
Quepha1 votesYes, yes really. People who are less experienced with a champion will generally have a lower win rate than people who "main" the champion. If the champion has a high popularity the people who "main" the champion make up a smaller percentage of the total games played with that champion. This drives the overall win rate of the champion down.
People think that, they say that happens, but in reality drastic play rate changes have very little effect on a champion's winrate. Whether that's because just as much of the large play rate is from past fans returning to their champion, mains returning to the game now that their champ is strong again, or some other factor is up for debate, but this hypothesis does not line up with actual observations.
Randomonium1 votesProof please.
Quepha1 votesProof please.
http://www.lolking.net/champions average winrate of the top 10 most popular champions: 49.75% That is nowhere near being far enough away from the average of 50% to constitute a trend.
average winrate of the top 10 LEAST popular champions: 48.31% This is less than 50% (although still not by a very significant amount) when these champions should, by the given hypothesis, be played almost exclusively by players who main the champion and know their champion inside and out.
Popularity of the top 10 highest winrate champions: 5, 20.1, 9.7, 10.7, 5.2, 9.5, 12.3, 4.9, 6.9, 3.3 These values are all over the place. Some champions are very unpopular, some have average popularity, and a few are very popular. There is no clear tendency towards any range of values.
Popularity of the top 10 lowest winrate champions: 1.2, 4.7, 20.6, 5.1, .7, 2.5, 3.2, 3.2, 3.7, 2.9 Most of these values are pretty low, although there is one very high outlier and a few others that approach average popularity (10 champions in a game out of 139 champions would be about 7.2 on average). With low play rates these champions should be played by knowledge-able mains that can succeed on the champion and have very few inexperienced players that drag the champion down.
This was a limited analysis of available statistics, but it seems to be 4 reasonable places to look for evidence that all either don't support the hypothesis or outright contradict it.
I think this hypothesis gained ground due to repeatedly seeing spikes in popularity accompanied by drops in winrate due to champion reworks, with a reversal of this trend on the following patch. But this phenomenon is almost always following a rework which has a much more obvious and logical source: players do not know how to play the champion yet and are still learning combined with that 'new champion' smell that attracts a lot of players. They are both separate trends that don't really affect each other even if they are caused by the same event.
Randomonium1 votesLooking at popularity and win rates between different champions has nothing to do with this discussion. What needs to be looked at is how win rates change for a SPECIFIC champion as popularity changes.
If we look at patch 6.5 for example, Vi, Talon, Soraka, Sejuani, Pantheon and Jarvan all saw this inverse relationship between popularity and win rate after they were buffed/nerfed.
Vi got buffed in patch 6.5. We saw win rate and play rate increase during this patch (as to be expected.) In patch 6.6 and 6.7 popularity dropped but win rate continued to climb dispite no changes to the champion. http://champion.gg/champion/Vi/Jungle
Talon got nerfed in 6.5. We saw play rate and win rate drop in 6.5. However, in patch 6.6 and 6.7 we saw win rate increase while play rate continued to decrease. http://champion.gg/champion/Talon/Middle
Sejuani got buffed in 6.5 and we saw both play rate and win rate increase in 6.5. However, in 6.6 play rate remained constant and win rate dropped. This caused play rate to drop in 6.7 which led to an increase in win rate with no changes to Sejuani's kit. http://champion.gg/champion/Sejuani/Jungle
Pantheon got nerfed in 6.5 and we saw a decrease in play rate and win rate. His win rate stabilized in 6.6 but his play rate continued to fall, causing his win rate to increase in 6.7. http://champion.gg/champion/Pantheon/Top
Jarvan got buffed in 6.5. We saw a huge increase in his play rate in 6.5 (along with an increase in win rate as to be expected) but then play rate dropped again in patch 6.6. However, we see his win rate continues to increase. More proof that win rate is inversely proportional to win rate is that his play rate between patch 6.6 and 6.7 remained the same and his win rate also remained the same. http://champion.gg/champion/JarvanIV/Jungle
That's just looking at ONE patch. Riot has also said that they don't look at just win rates for this exact reason and they put a lot more stock in what the win rate of a champion is among the champion "mains."
So yes, champion popularity and win rate are related.
Quepha1 votesHer biggest jump in winrate was also accompanied by her biggest jump in play rate. This seems like clear evidence that play rate and win rate are unrelated.
Winrate is all over the place while play rate steadily decreases. This seems like play rate and win rate are unrelated.
Biggest winrate boost accompanied by biggest popularity boost, then winrate decreased with playrate, then winrate increased as playrate decreased. This seems like play rate and win rate are unrelated.
Play rate and win rate are both generally decreasing steadily. This seems to run contrary to your hypothesis
Biggest winrate boost accompanied by biggest popularity boost.
How are these the best examples you can come up with? At this point I am more certain than ever that play rate has no clear correlation with win rate, other factors are clearly much more important.
Randomonium1 votesObviously I didn't explain myself well enough. When a champion gets a buff/nerf we expect win rate and play rate to increase/decrease as well. Look at the play rate and win rate the patches after the buff/nerf and you will see an inverse relationship between play rate and win rate. This is what I'm talking about. With no change to the champion's kit we see that win rate increases as play rate decreases and vice versa.
Quepha1 votesObviously I didn't explain myself well enough. When a champion gets a buff/nerf we expect win rate and play rate to increase/decrease as well. Look at the play rate and win rate the patches after the buff/nerf and you will see an inverse relationship between play rate and win rate. This is what I'm talking about. With no change to the champion's kit we see that win rate increases as play rate decreases and vice versa.
You are drawing a conclusion from very flimsy data because it is focused on very short periods of time following very specific events that you admit contribute the vast majority of the difference in values over that period of time. There is no useful correlation between playrate and winrate.
Randomonium1 votesFeel free to check my work with other patches. I chose a random patch and random champions that got buffed/nerfed in that patch and the shoe fit for each one. I chose each of those champions because we don't usually see play rate changes in champions unless there is a buff/nerf to the champion or the meta shifts. Its very suspicious that in each instance the data supports my hypothesis.
If you do more research you'll see the Reds hint at this as well. I'm not going to spend the time going through every single patch just to prove an obvious point. Take it or leave it but I've proven my point with a lot more evidence than you have.
Good day.
Wolfram Oxford1 votes52%+ is considered good, while some champions that have win rates below 48% are either mechanically difficult, weak, or both. By the way, 55% win rate is insane considering that this would be across a large variety of players varying in skill level. Although you may see 5% as small, it is actually pretty big when viewed as an overall win rate.
o Maui o1 votes55% is an insanely high win rate. Remember that this is over an enormous number of games, and that a 55% win rate over 1,000 games means that the champion won 100 games more than they lost. That's a massive difference, especially if you consider that a champ might easily be played that many times in just one hour (unless it's Galio in which case it might take a year). That big of a difference means that something is going on, especially for a more popular champions where the sample size is getting into the hundreds of thousands of games.
But I do agree that complaining over a 49/51% winrate is a bit excessive.
