How It Works
The methodology behind the ratings.
Two numbers, not one
The site carries two ratings and they are not the same measure, so every table names which one it is showing.
Elo is a skill rating on a Glicko scale. It updates match by match and carries across seasons rather than resetting, so a player starts a new season where the last one left them. Beside it sits a rating deviation, which is how much confidence to put on the number. Because it measures level rather than counting points, it is what the rankings fall back on before a season has any football in it.
Z-Score compares a player's points per game with the average for their position, on a scale where 1500 is that position average. Above 1600 is a strong performer and 1700+ is elite territory. It describes the season in progress and nothing else, so before the first gameweek it is held back rather than shown as current.
Tiers are bands of the players who have a measurement, not marks on the rating scale. Elite is the top 5% of 339 rated players, and each band below it takes the next slice. A tier is a rougher form of the rank printed beside it, so it can never claim more than that rank already does, and a player with no completed match behind them carries no tier at all.
The bars beside an Elo say how much football is behind it. One bar is a rating built on a handful of matches, which can move a long way once games are played, and three bars is close to a full season of evidence behind the number. A rating with a single appearance behind it is still ranked on it, because leaving a player out of the ranking is its own claim about them, and the bars are there so that rating does not read like one built on a full season. What the bars cannot tell you is whether there is a rating at all. Where this data carries none the player sits under the board with no rank, and some of those players do have Premier League minutes behind them.
Form trends
The form arrows (▲▲▲ to ▼▼▼) compare a player's recent output against their medium-term average. This helps identify who's heating up and who's cooling off before it shows in the headline stats.
Regression flags
Some players get lucky. Others get unlucky. Our system identifies players whose output doesn't match their underlying performance, helping you buy low and sell high.
Edge indicators
Special badges highlight players with distinct advantages:
Key metrics
How the numbers get here
This site is static. It is built from a data snapshot committed alongside the code, so what you are reading is whatever that snapshot said when the site was last built. Nothing on the page is fetched live.
The snapshot is regenerated by hand and the site rebuilt from it. There is no automatic job behind that, so we do not promise a refresh window. The data in this build was taken on 29 August 2026.
Every date and time on the site is UK time, written in the Europe/London zone rather than in whatever zone the machine that built the page happened to be set to. Kickoffs in the strip at the top therefore stay right on both sides of the clock change.