Studying the Correlation between Champion Pick Rates and Win Rates

With the arrival of Patch 5.4, discussion and controversy on champ balance blew up on various forums. A lot of it centered around junglers. Discussions over which champs were strong, if the top tier junglers were too strong for any other junglers to be competitive, and what Riot’s balance goals should be. 

Discussion shifted to win rates. I was reading a discussion on Reddit, the champion being talked about was Sejuani. A guy claimed Sejuani was useless, unplayable. How could Riot leave this champ so useless? Another guy brought up that Sejuani has a 54% win rate in ranked solo queue, the highest win rate of any jungler and ranking her as the 5th best champion in the game. The guy simply replied “Sejuani has a lower pick rate so only Sejuani mains play her, that’s why she has a high win rate. Everybody knows that.” He went say that while Jarvan (52.29%WR) has a lower win rate than Sejauni, he was much stronger than her but the win rate was misleading because since Jarvan is popular in LCS, bad players will see Jarvan and decide he must be overpowered, they go right into ranked and lose. I saw comments and upvotes flooding in to agree.  

League of Legends boasts 27 million players. I started looking through champion data from the past week, data compiled from over 1 million games. I was a little taken back that popular opinion is happy to generalize the 40,000 people who played Sejuani last week as experienced and cunning masters of the champion, while brushing off Jarvan’s 85,000 players as inexperienced noobs jumping on the hype train.

How can you generalize that? I really found the situation laughable. I knew previous research and from word of mouth that there are champs at every end of the spectrum high win rate and low win rate (Urgot comes to mind). Is there really a correlation? Can you say if pick rate has an impact on win rate? I started studying the numbers.

I gathered the pick rates and win rates from games for the last week, adding up to over 6 million picks total. 

On the graph below, each dot represents a champion. The farther right, the higher the win rate, and the farther upward they are, the higher number of picks. There are 121 champs now so it gets kind of messy. For that reason I decided to sort the champions. 

Win Rate: Anything at 50% or higher is categorized as “high win rate”, anything lower as “low win rate”

Popularity: The average pick number for each champ would have been 50,000. All champs above that are categorized as popular, all champions with a below average pick rate are categorized as unpopular.

High resolution mirror

High Win Rate champions are almost perfectly split between popular and unpopular champion picks. Meanwhile, champions with poor win rates are overwhelmingly unpopular picks. 

I can say with confidence that not only does the data that a champion with a low pick rate is not more likely to have a high win rate, it proves the opposite. A high pick rate champion is much more likely to have a high win rate than an unpopular champion. 

I believe that with so many tools to check champion stats and find guides, with League of Legends eSports viewership at an all time high, and with the popularity of social media and communities such as Reddit, more and more players are staying informed about the game. People are more likely to flock towards the stronger champions.

High resolution mirror

The top 10 most popular champions average a 51% win rate, with 7 of the 10 maintaining a win rate above 50%. Not only are most of these champions very strong, but these champions are also have some of the most well done skillsets, along with the best back story and theme, that make them truly fun to play. 

A lot of really strong champions here, strong on the current patch and able to make a strong impact in games. 

A trend I also noticed, champions who are popular but have a below average win rate often have a high skill cap. Thresh, Lee Sin, Riven, Vayne, and Zed are all in this category.

High resolution mirror

The 10 least picked champions have an average of a 46.6% win rate, with only one of the 10 (Poppy) having a win rate above 50%. I don’t think I can say anything about these champions other than that they are in a poor state. 

 Closing

For every low pick rate champ that is performing well, there are two performing poorly. The myth that low pick rate champions have a higher win rate is just that, a myth.