I'm curious how hero win rates and picks differ by region and rank in Dota 2. What data would make those comparisons useful, and how much should I read into the differences?
How would you compare Dota 2 metas across regions and ranks?
Answers
Start by splitting the data by patch, region and rank bracket. A hero's overall win rate can hide a lot if the hero is picked much more often in one bracket than another.
Look at pick rate alongside win rate, and include the number of matches. A big swing based on a small sample is less convincing than a smaller difference across lots of games. Comparing similar periods matters too, since a balance patch can change the picture quickly.
Reply to this answer
Don't treat a higher win rate as proof that a region plays a hero better. The hero might be picked into different matchups, played in different roles, or selected mostly by specialists. The stats are a good starting point for questions, not a full explanation on their own.
Reply to this answer
For Dota, I'd focus on hero picks rather than map picks. If you want to study map play, you could compare things like objective timing or where fights happen, provided your data actually includes those details.
Reply to this answer
Add your answer
Answers are reviewed before publication.
DLCmapMittens:
Would you compare every rank bracket separately, or combine some of them?
critBudget_6:
Separately where there's enough data. Combining neighboring brackets can help with small samples, but label the grouping so the comparison stays clear.