everystat
How the six-axis ratings work
Every ranking on everystat is six axes on a 1–100 scale. This page says exactly what goes into each axis and what we adjust — so a rank can be cited for what it is, and not for what it isn't.
What a rating is
A rating is a percentile inside one ranking (role × season). 100 means top of that group; 50 means the middle. Ties share the midpoint.
Each axis is the mean of its metrics' percentiles. The overall score is the mean of the six axes.
Ratings are not comparable across rankings or seasons — the group changes. Compare only inside the same page.
The six axes and their metrics
| Axis | Metrics (laners) | Jungle differs |
|---|---|---|
| Laning | Gold, XP and CS difference at 15:00 vs the lane opponent; plus solo kills conceded before 15:00 where 80% or more of the ranking has that data | Same |
| Teamfight | Damage share, kill participation, damage mitigated per minute | Same |
| Mechanics | Solo kills per game, damage per minute, CS per minute | Damage share, damage per minute, solo kills per game, CS per minute capped at 7 (above that, all tie) |
| Macro | Vision score per minute, wards cleared per minute, control wards bought | Adds CS, gold and XP difference at 15:00 — counter-jungling and early tempo |
| Consistency | Win rate, deaths per game (fewer is better) | Same |
| Champion pool | Distinct champions played in the season | Same |
Adjustments — every one of them
- Few games, pulled toward 50: rating = (percentile × games + 50 × k) ÷ (games + k), with k = min(40, 0.4 × the most games anyone in the group played), never below 5. A 5-game wonder cannot top a 40-game season.
- International events (MSI, Worlds, EWC) are measured against the same-year, same-role distribution of the four major leagues (LCK, LPL, LEC, LCS) — not against the 13–16 players at the event. Otherwise a mid-table player looks elite because play-in teams sit below them. Needs at least 20 major-league players. The rank order inside the event is unchanged; only the values are.
- Strength of schedule, international events only: each team's opponent Elo becomes a z-score, and axes that measure results against opponents shift by 6 percentile points per standard deviation — laning fully, consistency fully, mechanics one third (solo kills only). Teamfight, macro and champion pool are not shifted: they measure share of one's own team or absolute counts.
Limits — read these before citing a rank
- A rank is not a raw metric. Because an axis blends metrics and adjustments, a player can be 1st on the laning axis while 31st on gold difference at 15. Every ranking page lists the raw metrics — cite those when precision matters.
- Solo kills conceded is reconstructed from match feeds and is a lower bound. It has its own method page.
- Year-wide rankings that combine leagues use only the metrics every league in the group can supply. Where minor leagues lack solo-kill coverage, laning uses three metrics for everyone in that ranking.
- Sources: collected professional match records and Riot Games' official Match LiveStats. Not endorsed by Riot Games.
How to cite
"Laning 87 (3rd of 40, LCK Mid 2026) — everystat, percentile within the ranking; raw gold difference at 15: +312." Name the ranking, give the rank with the group size, and add the raw metric.