Every player on fplwire carries an expected points (xP) number for each of the next few gameweeks. This guide walks through what goes into that number, with real figures from the GW10 2026/27 run, so you can read an xP score the way the model actually built it, and know when to override it with your own judgement.
What xP actually adds up
In fplwire’s model, expected points for a single match is the sum of five parts:
See what this means for your own team
Open fplwire for your teamFree. No account, no FPL password.- Appearance points: 2 points multiplied by the chance the player plays 60 minutes or more.
- Attacking returns: the goals and assists the model expects, converted into FPL points for the player’s position.
- Clean sheet points: 4 points for goalkeepers and defenders, 1 for midfielders, following FPL’s own scoring rules.
- Goalkeeper save points: worth about 1 point a game for a goalkeeper.
- Bonus points: the extra points awarded to the best performers in a match.
Add those five together and you get the xP figure shown in the app. Take Haaland’s GW10 fixture away at Nottingham Forest: the fplwire database (model run 1 Nov 2026) has him at 8.15 xP, built mostly from his appearance points plus a large attacking returns component given his scoring record. Gabriel at home to Hull is 5.22 xP, a mix of appearance points, clean sheet points and a smaller attacking contribution. Raya, also at home to Hull, sits at 4.91, leaning heavily on clean sheet and save points rather than goals.
How the “chance of playing” is set
Every part of a prediction is multiplied by the chance the player plays, so this number matters more than any other. Since 3 November 2026 fplwire sets it from two things:
- Recent minutes: how much of each of his team’s last four gameweeks the player actually played, with the most recent week counting most. A player who played every minute of all four is treated as about 95% likely to play; one who sat on the bench for all four drops to 5%.
- FPL’s availability flag: a doubtful player is capped at 50% and an injured or suspended one at 5%, however many minutes he played before.
We changed this after checking our own predictions against real results for GW6 to GW9. Before, every player FPL listed as available was treated as 90% likely to play, yet 37% of them did not play at all in GW6 to GW8. Predictions were too high on average, and bench players looked like starters. With recent minutes, the average prediction now sits close to what players really score.
The flag still matters. Saka was at 0.24 xP for GW10 in the 1 November run because FPL flags him injured, which caps his chance of playing at 5% and takes almost his whole projection with it.
How attack and defence strength are measured
For the attacking returns and clean sheet parts, the model compares two team-level numbers, according to fplwire’s model code. Team attack is a side’s goals per game measured against the league average. Opponent defence is the goals that opponent concedes per game, also measured against the league average. Both figures are capped between half and double the league average, so one unusually high or low scoring game cannot distort the rest of the model.
Home and away adjustment
On top of that, home teams get roughly an 8% boost to attacking returns and away teams take roughly a 5% cut, per the model code. That is one reason Haaland’s away trip to Nottingham Forest (8.15 xP) sits below his home fixture a week later against Fulham (8.64 xP in GW11), despite Forest and Fulham being different defensive propositions in their own right. Fixture difficulty and venue both move the number, not just the opponent’s reputation.
Clean sheet probability
Clean sheet chance starts at about 35% against an average attack, rises against weak attacks and falls against strong ones, and the model code never lets it drop below 5% or climb above 65%. That is the mechanism behind Gabriel’s run of clean sheet-driven defender xP: 5.22 at home to Hull, 5.1 away at Newcastle in GW11, 4.5 at home to Manchester City in GW12. Notice the Manchester City fixture produces his lowest number of the five gameweeks shown, because that opponent’s attack pulls his clean sheet chance down even with home advantage.
Why form takes time to kick in early season
Form, in the fplwire model, is a player’s points per game measured against the average for their position. The catch is that early in a season there is not much evidence to go on. According to fplwire’s development notes (24 Aug 2026), the model starts each player’s form estimate from their price relative to others in the same position, then leans more on actual results as matches accumulate: after one match played, the estimate is still roughly eight ninths price-based, and by around ten matches real results carry more weight than price. The notes record what happened before this was added, in August 2026: the model ranked six forwards above Haaland despite all of them having scored zero points, purely because price-based form was missing. That blend of price and results is now built in specifically to stop a small early sample from producing a chaotic ranking.
The six gameweek cutoff and the 20 point cap
fplwire scores every Premier League player for each of the next six gameweeks using the full model above. Beyond that, gameweeks out to GW38 still get a number, but the app labels it a rough estimate, according to fplwire’s model code. Treat those far-ahead numbers as a general guide to fixture swing rather than a precise prediction.
There is also a ceiling: no player can be given more than 20 expected points in a single gameweek, even in a gameweek with two matches, per the model code. This stops an outlier combination of two very favourable fixtures from producing an unrealistic headline number.
What the model is not told
This is the part that matters most for reading xP sensibly. According to fplwire’s model code (checked 2 Nov 2026), four things are not currently built in:
- Defensive contribution points: the 2 points awarded when a defender records 10 clearances, blocks, interceptions and tackles in a match, or a midfielder or forward records 12 including ball recoveries, a rule that continues in 2026/27 according to premierleague.com’s Fantasy rules article. The model does not count these as a separate part, but they still reach the number indirectly: form is points per game, and those points include defensive contributions. When fplwire tested adding them as a separate part on GW6 to GW9, predictions did not get better, and players who earn them regularly did not beat their xP.
- Penalty and set piece duty: the model does not know who takes penalties or corners for a team, so it cannot weight a known penalty taker’s attacking returns any higher than a teammate’s.
- Expected goals and shot quality: the attacking returns figure is built from actual goals scored, not the quality of chances created, so a player on a hot or cold scoring streak will show that streak in their xP rather than an underlying shot-quality read.
- Rotation before it happens: the model sees a player lose his place only after he has missed matches. A planned rest, a cup week or a new manager’s first team sheet is not in the number until the minutes show it.
None of this makes the xP number wrong, it is accounting for exactly what it says it accounts for. It just means a manager who knows a team’s penalty taker, or who has seen a squad rotate in cup weeks, knows something the model does not.
Where xP shows up in the app
These xP numbers are not just a standalone figure. According to fplwire, they drive player rankings on My Team, the Pick my best XI tool, the transfer planner, and they also feed into the AI recommendations as one input among others. Understanding what is and is not in the number helps when any of those tools lean on it.
What to do with this
- Check the player’s FPL status and recent minutes before trusting his xP. A doubtful flag caps his chance of playing at 50%, and a player who has been on the bench will show a low number even if he is fit.
- Treat gameweeks beyond the next six as a rough estimate, useful for spotting a run of good or bad fixtures, not for pinning down an exact score.
- If you know who takes penalties or set pieces at a club, weigh that yourself. The model has no penalty data built in, so it will not reflect it.
- Do not add defensive contribution points on top of xP in your head: they are already reflected through each player’s points per game.
- Before each deadline, check team news for rotation the model cannot see yet: it learns from minutes already played, not from what a manager plans next.
See what this means for your own team
fplwire reads your live FPL team and shows you your captain options, transfer targets, chip timing and mini league standings in one place. Free, and no account needed.