How the Expected Points Explorer works

Expected points is a statistical estimate of how many FPL points a player is likely to score in a given gameweek. Unlike raw form — which tells you what happened last week — xP is forward-looking. It combines a player's underlying output with the difficulty of their upcoming fixture to produce a single number that represents their scoring potential.

A high xP doesn't guarantee a return. A player projected at 7.2 xP might score 2 points or 15 points on any given Saturday. What the number tells you is that across many equivalent fixtures with a similar player in similar form, the average outcome lands around 7.2. Over a full 38-gameweek season, consistently picking the highest xP options outperforms chasing last week's big scorer or reacting to newspaper headlines.

Live betting odds are the single most predictive input. The markets price in team form, injury news, managerial tactics, and head-to-head history faster than any model can — and they're available for every match in the upcoming round. The model pulls match-result odds, clean sheet odds, and over/under goal totals and uses them as the foundation for every player's projection.

On top of the odds, each player's Understat non-penalty xG and xA from the trailing season is blended in via a Bayesian prior. This smooths out noise: a striker who missed three sitters in a single week doesn't get punished in the model, and one who scored a lucky long-ranger doesn't get rewarded beyond what their underlying numbers support. The prior also means early-season projections lean on last season's rates until enough new data accumulates.

A home-advantage multiplier is applied to all players in home fixtures, consistent with the roughly 10–15% uplift in attacking xG that playing at home produces across European football. Finally, every player's xP is scaled by their starts probability — the fraction of recent gameweeks they actually started rather than appeared as a substitute. A player who starts around half their games has their projection halved accordingly, making rotation risk visible at a glance. The full technical methodology is covered in the Squad Optimizer guide.

Live betting odds are only published for the next confirmed gameweek. Beyond that, the model switches to the FPL Fixture Difficulty Rating as a proxy, scaling each player's next-GW xP up or down depending on how hard their future fixtures look. The scaling factors are:

A player projecting 6.0 xP in GW+1 and facing an FDR 1 fixture the following week would be projected at 7.5 xP for that round (6.0 × 1.25). Double gameweeks are handled by projecting each fixture separately and summing them, so a double will produce a noticeably larger number than a single-fixture week.

The important caveat is that anything beyond GW+1 is an estimate, not a live-odds projection. The GW+2 to GW+5 columns are best used for directional planning — deciding which players to bring in before a fixture swing, timing chip usage, or identifying who has the kindest run of games over the next month — rather than as precise point predictions.

Each row is a player. The GW columns show their projected xP for each upcoming gameweek — a missing value means they have a blank that week, while a double gameweek produces a visibly higher number because both fixtures are summed. The Total column adds up xP across all the gameweeks you've selected, and the Avg column shows the per-game average excluding blank weeks. Avg is generally the better metric for comparing players who face different numbers of blanks, since Total will naturally be lower for a player who has one fewer fixture.

The GW range selector at the top (1 through 5) controls how many gameweeks you're projecting over. Narrow down by position using the GKP/DEF/MID/FWD buttons, filter by team using the team dropdown, or search by player name. Clicking any column header sorts the table by that column, so finding the highest-xP midfielder over the next three gameweeks takes about three clicks.

Each xP cell is colour-coded so you can read the table at a glance without processing every number individually. The thresholds are:

Only the next gameweek uses live odds. Everything from GW+2 onwards is scaled from FDR, which means the further out you look, the more directional and less precise the projection becomes. Treat the outer columns as a fixture-difficulty guide rather than a point forecast.

Goalkeepers are the noisiest position to project. Clean sheet probability accounts for most of a goalkeeper's xP, but save points and penalty saves are close to random from a statistical standpoint. The model does its best, but two similarly-priced goalkeepers with similar clean sheet odds can have very different actual returns on any given week.

Starts probability is calculated from recent history and takes a few gameweeks to catch up when something structural changes — a new signing, a manager switching formations, or an injury to a rival for the position. If you know a player's role has shifted recently, factor that in manually. Similarly, the table doesn't automatically optimise around blank and double patterns. Use the Total and Avg columns together to compare players who face different fixture schedules over your selected range.