PERFECT XI
Home / Guides / The Data
THE DATA

INSIDE THE RATINGS DATABASE

Every spin of the wheel draws from a hand-balanced database of 2866 players. Here is how it was built, what the numbers mean, and the exact top of the pile.

Written by the Perfect XI developer · Updated 8 July 2026

The database at a glance

Perfect XI runs on a hand-built ratings database covering 3,008 unique players across 51 clubs and 714 club-seasons of the Premier League era (1992–2026). Every player who appears on the draft wheel comes from this dataset — there are no generated or placeholder names.

The average rating across the whole database is 77.0, and the distribution is deliberately top-heavy in its scarcity: only 74 players are rated 90 or above (2.5% of the database), with another 219 in the 85–89 band. When a 90+ card lands on your spin, it is genuinely rare — and the maths of squad-building reflects that.

Regular vs prime ratings

Each player carries two numbers: a season rating for the specific club-season the wheel lands on, and a prime rating representing their peak in that shirt. The game's default mode drafts on prime ratings, which is why a 2002 spin can hand you the best version of a player whose career peaked later.

This is also why the same player can appear at different strengths for different clubs. Alan Shearer's Blackburn and Newcastle entries both peak at 94 — but plenty of players are meaningfully stronger in one shirt than another, and knowing which era to hope for is part of the skill.

The ten highest-rated players in the game

Cristiano Ronaldo99Man United2002
Thierry Henry97Arsenal2003
Erling Haaland95Man City2021
Steven Gerrard94Liverpool2008
Mohamed Salah94Liverpool2019
Kevin De Bruyne94Man City2017
Alan Shearer94Newcastle1996
Fernando Torres93Liverpool2008
Peter Schmeichel93Man United1998
Wayne Rooney93Man United2003

Cristiano Ronaldo's Manchester United entry tops the database at 99 — the only 99 in the game — with Thierry Henry's Invincibles-era Arsenal at 97. Note how the top ten spans all four decades: the database is balanced so that a 1990s draft can compete with a 2020s one. For the full club-by-club picture, see every club's highest-rated player.

How ratings were decided

Ratings are a fan interpretation built for gameplay, derived from a public football dataset and then balanced by hand for the draft format. Three principles guided the pass: era fairness (each decade needed championship-calibre spines, or era-locked drafts would be unplayable), positional scarcity (elite goalkeepers and defenders are rarer than forwards, mirroring the real market), and within-club honesty (a club's all-time hierarchy should feel right to that club's fans).

They are not official statistics and do not claim to be — see the disclaimer in the footer. They exist to make the draft fair, tense and arguable, which is the point of the game.

What this means for your draft

Two practical consequences. First, because the average card is a 76–77, an all-90s fantasy XI is mathematically off the table — a realistic elite squad averages in the high 80s, and the match engine is tuned around that. Second, scarcity means positional patience pays: passing on an 84 forward to hold out for a defender upgrade is often correct, because the forward pool is deeper. The best XI of every era shows what the ceiling actually looks like.

▶ Play Perfect XI free

How ratings are calibrated (2026 rebuild)

The rating scale was rebuilt in August 2026 to be anchored on evidence rather than reputation. A player’s peak rating is now tied to what he actually won and produced.

Every rating of 90 or above now maps to a specific honour: a PFA Players’ Player of the Year, a Premier League Player of the Season, a Golden Boot, or a Ballon d’Or placing. Players who previously sat at 90+ on reputation alone were moved down into the 85–89 band, which cut the 90+ group from 74 players to 26.

The rebuild also corrected a bias that affects most fan-made databases. Marquee signings whose fee exceeded their output were adjusted down, while defensive midfielders, full-backs and creators — roles that goals-and-assists systems undervalue — were adjusted up.

One club per player per season

The database also carried players in squads they did not play for in that particular season, usually attributed a year early or a year late relative to when they actually signed. Every case was resolved by tracing each player’s career run and assigning the season to the club he genuinely played for, which corrected 471 entries.