I like this idea! But any type of win rating percentage would be theoretical at best. imo at least. I mean come on a 75% win NS that goes boots of travel and Codex(Dagon) will lose agains a 30% TideHunter that decided to form a carry/tank build.
I came here to post exactly that.
I completely agree that the current system doesn't really tell us much.
The deviation from average is a good idea. Or you could make an index; hero matchup/avg matchup, showing with which factor the specific matchup is better or worse than average.
Pu12e's argument that how a hero fares against another hero is build-dependent goes for any realistic matchup analysis.
Чтобы оставить комментарий, пожалуйста, войдите в систему.
Heroes matchup stats as it is right now is hard to analyze correctly as heroes with an overall low win rate will typically perform poorly against all heroes and be at the top for each hero matchup. For instance, Meepo is cited as the best matchup for each hero.
What could be interesting is display how a hero perform against the average of the other hero. For example, if I play Wisp and have a 55% winrate against Broodmother but her overall winrate is 40%, Wisp is not a got choice against her as its winrate is below the average 60%.
What could be displayed instead and would be easier to analyze is the difference from the average. In the previous example, Wisp would show a -5% against Brood. While it is not a perfect solution, it would be much easier to see against which heroes a certain hero is better than the average.