Rankings

How the rankings work

Every skin starts at 1500. Every battle moves two numbers. Everything else is detail, and the details are fun.

Placements, basically

Every one of the ~1,900 skins starts at 1500, with a big uncertainty band of ± 350. Think of a fresh account in ranked: the system has no idea how good it is yet, so placement games swing hard. A skin's first battles work exactly like that.

As evidence piles up, the band tightens, down to a floor of ± 60. That ± number is shown everywhere on the rankings, and it means what it says: a skin at 1480 ± 90 could plausibly sit anywhere in that range. The ranking is honest about what it knows.

Every battle moves both skins

When you pick a winner in a battle, the winner takes rating points and the loser pays them. How many depends on two things:

How surprising the result is. A favorite beating an underdog tells us almost nothing, so the numbers barely move. An upset is real information, and the system rewards it accordingly.

How settled each skin is. A brand-new skin can swing by up to 64 points in a single battle. A skin with hundreds of battles barely twitches, around 16 at most. Fresh skins find their level fast; established ones do not get knocked around by one weird Tuesday.

A worked example: the upset

Elementalist Lux sits at 1620, mostly settled after dozens of battles (± 70). Pool Party Graves just hit the arena at 1430, still wobbly (± 300). The math expects Lux to win about 3 times in 4. You pick Graves.

Pool Party Graves

1430 1472 (+42)

New skin, surprising win: maximum learning, big jump.

Elementalist Lux

1620 1607 (−13)

Settled skin: one upset dents her, it does not define her.

If Lux had won instead, she would have gained about 4 points and Graves would have dropped about 14. Expected results barely move the needle. Upsets are where rankings are made.

Your vote has weight. Literally.

Battles you fight as a guest count at half weight. Sign in and every pick lands at full strength.

Here is the part we are proud of: when a guest creates an account, their old votes are not stuck at half weight forever. The system keeps every battle ever fought and periodically re-reads the whole history, so your past picks get upgraded to full weight retroactively. Your week of anonymous swiping was not wasted. It was an investment.

Why some lists say "still calibrating"

A ranking earns full confidence once the typical skin on it has 10 battles. Below that, we show the list anyway, banner and all, because thin data is not something to hide. It is something to fix, and every battle you fight fixes it a little.

Skins with very few battles also wear an "early ranking" tag on their own pages. Same idea: the number is real, the confidence is not there yet.

The recount

The instant +42s and −13s you see after each pick are the live scoreboard, tuned to answer you back immediately. Every few hundred battles, the system also does something slower and more serious: it replays every battle ever recorded from scratch and rebuilds the entire ranking from the full history.

That recount is the official result. It irons out streaks, lucky matchups, and ordering quirks that live updates can drift on. It is also why a skin's rating can shift slightly even on a day nobody battled it: the room got smarter about everyone.

For the stats nerds: the live updates are Elo-style, the recount is a Bradley-Terry model fit over the raw battle log. Elo experience, Bradley-Terry truth.

The model is hungry. Feed it.

Battle nowSee the rankings