Mei talent statistics: how to read the results
Start with the game count, check the talent choices, then bring the draft and the eye-test back into the decision. This guide uses Mei's available replay sample; it does not declare a best build from percentages alone.
Statistics snapshot retrieved . Results may change as replays arrive.
What does the Mei sample contain?
This sample uses all modes and all ratings, replay build 98348. Latest means the newest build in the available replay data, which may lag the live game. It is the same default selection used by the talent page.
Mei appears 21 times, with an overall win rate of 57.1%. These are contributed replay observations, not a random sample of all HotS players. Match the build and filters before comparing this result with another site.
Compare Mei's level 1 talent choices
At level 1, Ice Storm has the most observations: 11 games. Heavy Pack has a raw win rate of 100% from 2 wins in 2 games. Popularity and win rate answer different questions; neither establishes the right choice for your next draft.
| Talent | Wins / games | Win rate | Pick rate |
|---|---|---|---|
| Ice Storm | 6 / 11 | 54.5% | 57.9% |
| Heat Transfer | 3 / 6 | 50% | 31.6% |
| Heavy Pack | 2 / 2 | 100% | 10.5% |
How much would one more result change it?
One additional loss with Heavy Pack would change 2 wins in 2 games (100%) to 2 in 3 (66.7%). This is a hypothetical sensitivity check, not a prediction or an additional observed match.
Heavy Pack has fewer than 10 games, so it does not qualify for MercStats' highest-win-rate talent highlight. A large percentage can rest on very little evidence.
Players, teammates, opponents, maps and drafts all affect the result. More observations help, but these games do not isolate the talent's effect.
Check which games are in the denominator
The sample has 21 Mei appearances and 19 appearances with a recognized level 1 choice. Talent win rate divides wins with that choice by games with that choice. Pick rate divides its games by the recognized choices at that level.
Ice Storm's 57.9% pick rate means 11 of those 19 recognized level 1 observations. It does not mean 57.9% of everyone playing HotS chose it.
The hero and talent-tier counts differ. This aggregate does not establish a specific reason for that difference; do not assign a cause without inspecting the underlying evidence.
From individual Mei talents to a build
Choosing the highest-win-rate Mei talent at every level does not establish a win rate for that combined build. Talents can depend on earlier choices, and an exact complete path can have far fewer observations than its individual talents.
Read the talent descriptions and the explanations on the build pages. Published guide choices and observed replay results are separate evidence. Check whether the path fits your team's plan and the threats in the opposing draft.
Use the eye-test when playing Mei
Stats can point to a question to check in a replay. They do not always back up a useful play: creating space, enabling a teammate, or choosing a safer fight may not be clear in a talent win-rate table. The eye-test matters too, while the numbers can challenge an impression based on one memorable game.
Did the choice help you start a fight your team could follow, or protect a teammate who was under pressure?
Was your engage useful, or did it leave your damage dealers without protection?
For Mei, compare similar maps and drafts, inspect the moments when the choice mattered, and keep the sample size in view. If the replay and the percentage seem to disagree, investigate the context before declaring a winner.