Balancing based off of character winrate is actually a bad idea, and here’s why:

I can undestand why balancing based off of winrate sounds like a good idea, but is isn't, and I can explain:

Suppose we're got a cast of three characters, A, B, and C. A has a 50/50 matchup against itself (mirror), a 50/50 matchup against B, and a 100/0 matchup against C (so A is 50-50-100). B has a 50/50 matchup against A (already stated), a 50/50 matchup against itself (mirror), and a 0/100 matchup against C (so B is 50-50-0). From this you can deduce that C has a 0/100 matchup against A, a 100-0 matchup against B, and a 50/50 against itself (so C is 0-100-50).

If we assume all three characters are picked equally, that means A is going to win 67% of it's matches (75% if we exclude mirrors), B is going to win 33% of it's matches (25% if we exclude mirrors), and C is going to win 50% of it's matches. By this metric, C is the most balanced, even though it actually has no fair 50/50 matchups at all except for mirrors, which means it's really the least balanced. If it weren't for C, A and B would each be at 50%.

Unless you want to let your game become crummy rock-paper-scissors counterpicking, you need to balance characters off of their matchups and not their overall winrate.


FURTHERMORE, balancing based off of winrate inherently favors popular characters. A more popular character is going to have more mirror fights, which add both one win and one loss to its overall winrate, bringing it closer to 50/50. An unpopular character is going to have fewer fights against itself, and instead have fights against other characters, meaning that it won't get the same treatment.

I found some data off of imgur which shows what characters were the most used during the beta:

http://imgur.com/gallery/IsgO6

As you can see, in nearly every gamemode assassins made up 50% of the playerbase. Orochi was in front, and Peacekeeper was tied for second place with Raider and Warden. This is going to actively bring their win/loss ratio closer to 50/50 due to the increased odds of a mirror matchup. In other words, their win/loss is watered down by popularity, meaning that Raider will look less underpowered than he is, and Warden and Peacekeeper will look less strong than the other top tier characters (thus explaining why Lord and Conq have higher winrates).

Leave a Reply

Your email address will not be published. Required fields are marked *