Mastery and win rate, across every game the ladder holds
Mastery and win rate, across every game the ladder holds
Every seat on ranked, grouped by rank and by how much mastery the player had on that champion BEFORE the game started — not after, so wins earned that game cannot inflate the reading. Pick a rank to see its own curve; the population never needs a Riot ID.
Loading…
Every seat here used a mastery reading taken BEFORE that game — points a game itself awarded can never count toward its own result, so this cannot be measuring the reverse relationship.
This is not a promise about your next game. Players do not reach a mastery band in a random order, and a champion someone loses on tends to get dropped before it reaches the high bands — so the right-hand side of every curve is partly survivors, not only practice.
Which champions reward knowing them
The same measurement, one champion at a time: pooled win rate on a champion once someone has mastered it, minus their win rate the first time they pick it up. A big gap is a champion that punishes an unfamiliar pilot and rewards a practiced one — closer to a mechanical skill check. A small one wins about the same either way — closer to a stat check.
Loading…
A big gap is not the same claim as "this champion is strong" — it says the win rate MOVES with familiarity, not where it ends up. A champion with a rough floor and only an average ceiling can still top this list.
Nobody plays a champion to 300,000 mastery points by accident. Part of every gap here is players who kept losing on a champion dropping it long before they reached the top band — the same caveat the personal pool chart carries.
Which of your champions should you play more?
Point the same measurement at your own record. For every champion you play, this compares how you do on it with how you do on the rest of your pool — so it is about you, not about which champions are strong this patch. A 45% player and a 55% player can both be four points better than usual on their best champion.
It only ranks champions you have actually played. Similarity carries your record onto champions you have already touched, and we measured that it does not predict a champion you have never picked: without your own games in it, the neighbourhood ordered held-out results backwards as often as forwards. So this narrows what your record already says — it does not guess at the rest of the roster.
One lookup, and nothing is stored. It also opens the network below on your most-played champion.