Losing more on a new champion is normal, and it's cheap: learning one costs the average player less than one win. Most of the dip is gone within 20 games. Your win rate on the champion while you learn it is mostly luck, so don't judge a champion by its first 10 games.
Players win 45.8% of their first ranked game on a champion that's new to them, 5.0 points less than on their usual champions that month. We followed 677,151 players through 95 million ranked solo games since September 2025 and compared each player with themselves: their games on a new champion against their games that month on champions they already played.
Is it normal to lose more while learning a new champion?
Yes, and it costs you less than one win. You'll lose a little more in your first few games, but half of the dip is gone by your fifth game and nearly all of it by your fiftieth.
Over their first 20 games on a new champion, the average player wins 0.44 fewer games than on their usual champions, and 0.67 fewer over their first 50. The first game costs 4.9 to 5.3 points at every rank from Iron to Master.
Players drop a champion after losing on it, so the ones still playing it later are partly the ones it went well for (the dashed line). The solid line counts everyone who tried the champion, by giving more weight to the games of players whose early results usually make people quit.
The experience curve on stats sites, like a champion's Trends tab, compares different players and mostly holds up: it rises 5.6 points from the first game to the 51st to 60th, and the same players compared with themselves rise 4.8.
How many games before your win rate on a champion means anything?
More than you'll play while you decide whether you like it. Your first 10 games on a champion tell you almost nothing about the next 10, so don't drop a champion over a bad start, and don't keep one because of a lucky one.
Ten games are mostly luck. Even a player who wins exactly half their games goes 3-7 or 7-3 all the time, and the real difference between two of your champions is usually much smaller than that swing.
Players who started 7-3 on a champion won 53.6% of their next 10 games, and players who started 3-7 won 50.9%. Only the rare extremes say much: the 858 players who started 0-10 won 41.9% of their next 10, and the 3,799 who started 10-0 won 63.0%.
A player who wins exactly half their games over the long run goes 3-7 or worse, or 7-3 or better, in 34% of 10-game stretches. Over 100 games they still end at 45% or below, or 55% or above, 37% of the time. Two players 10 points apart over their first 100 games on a champion can expect to be only 2 points apart over their next 100.
Which champions take the longest to learn?
Rengar, Azir and Qiyana are the hardest to pick up: expect to lose a lot more in your first games, and to still be behind 50 games in. Seraphine, Sona and Amumu cost you nothing at all.
Riot's difficulty rating is only a rough guide. Yuumi is rated easy and is one of the harder champions to start on.
Players are still 7.7 points under their usual on Rengar in their 21st to 50th games, and 7.8 on Azir. Some of each gap is how strong the champion was this season, but taking that out barely changes the ranking: the two correlate at 0.91. Champions Riot rates 7 to 10 cost 6.2 points over the first five games and those rated 1 to 3 cost 3.2, but Yuumi, rated 2, costs 8.5.
Does one-tricking make a new champion harder?
Yes. If you mostly play one champion, a new one will feel much worse, because you're measuring it against your best. Expect about four times the dip of a player who spreads their games around, and for it to last longer.
Players who spread their games across many champions lose 3.1 points on their first game on a new one, and players whose main is 60% or more of their games that month lose 12.3. In games 6 to 20 the one-tricks are still 6.6 points under, against 0.9.
Should you give up on a champion after a bad start?
Most players do, usually within a few games, which is exactly when a new champion feels worst. A bad start doesn't mean much: players who push through 0-3 go on to win about half of their next ten games on it.
78% of new champions get four ranked games or fewer. Players play a champion again 49.5% of the time after losing their first game on it, and 63.1% after winning it. Players who start 0-3 and keep going win 50.9% of their next ten games, and players who start 3-0 win 54.3%. The ones who quit might have done worse than the ones who stayed, though the correction in the first chart suggests not by much.
What should you do when learning a new champion?
Give it 20 games before you judge it, and don't trust its win rate even then. Expect to lose a little more at first, and more if you're picking up Rengar, Azir or Qiyana, or coming off a one-trick. It will cost you less than one win, and most of what your record says along the way is luck.
We only see ranked solo queue, so some players will have tried the champion in normals, ARAM or the Practice Tool before their first ranked game on it.
