Injuries and fitness
The question is not just ‘is he flagged?’ It is whether the knock, illness or return to training changes his expected minutes.
How we make the call
What earns an alert, which sources deserve trust and where the fixture numbers come from.
Does it change the pick?
That is the test. An update makes the main feed when it could move a player’s minutes, role, set pieces or fixture value. General chatter stays out unless there is a credible source and a genuine FPL consequence.
Who we trust first
- 01Official club news and manager press conferences
- 02Competitive line-ups, substitutions and match reports
- 03Premier League and official FPL information
- 04Reliable national and specialist football reporters
- 05Rumours – labelled, handled carefully and usually left out
Selection and rotation
Starts, substitutions and a manager’s own words beat guesswork. Once competitive football begins, it carries the most weight.
Role and set pieces
Playing through the middle is different from hugging the touchline. Penalties, corners and free-kicks matter when the responsibility is real.
Transfers and discipline
A new signing, returning team-mate or ban can change a player’s route to the XI overnight.
Fixture and opponent
A moved match or a key opponent absence can change the FPL read, especially for clean sheets and captaincy.
How AI FPL xP works
The official FPL projection is the anchor, not the whole answer. We adjust it for a player’s price and position, expected minutes, set-piece role, team strength, venue and fixture difficulty.
New competitive minutes and underlying goal involvement feed in gradually. One hot result can move the number; it cannot take over the model on its own.
How the fixture call works
Each club’s attacking rate is its players’ combined expected goals per completed match. Its defensive rate is expected goals against per completed match, taken from official goalkeeper xG conceded data. Live and unfinished fixtures are left out.
Home model goals equal the square root of home xG multiplied by away xGA, with a 10% home lift. Away model goals reverse the inputs and apply a 10% reduction. Early-season figures are pulled towards the league average using three league-average matches, which stops one unusual result from controlling the forecast.
The two expected-goal estimates are rounded into the displayed score estimate. A Poisson distribution produces the separate home, draw and away probabilities. The scoreline is a planning estimate, not a promise. Confirmed line-ups, role changes and major absences still need to be considered separately.