Loading
A plain-language breakdown of how every prediction and metric on this site is actually calculated — the same explanations shown behind the (i) icon next to each metric, collected here in one place. Written for someone with no FPL-analytics background, not as a restatement of the underlying code.
Most of what follows leans on the same two guardrails, so they're worth understanding once rather than re-reading on every page:
Starter reliability
How much of the last few gameweeks this player has actually been on the pitch, out of 90 minutes each game (benched weeks count as 0). Below 70% it's flagged as a rotation risk, and every prediction for that player is scaled down to match -- a great underlying rate built on a couple of late cameos doesn't mean much if they might not start at all.
450-minute rule
Once a player has 450+ real minutes this season (about 5 full matches), their own current rate is trusted directly. Below that, a handful of minutes is noise, not a rate -- so the model uses their real, frozen rate from last season instead. Never the live match being predicted: a single shock result never gets treated as an established trend.
FDR (fixture difficulty)
The Premier League's own 1 (very easy) to 5 (very hard) difficulty rating for this specific fixture. Known before a ball is kicked, so it's safe to use for predicting a match that hasn't happened yet -- unlike a stat from the same game being predicted.
One more rule that applies everywhere on this site, not just predictions: nothing is ever calculated from the live/in-progress match it's predicting. A stat only becomes usable input once its own match has actually finished — using a match's own result to "predict" that same match would be circular.
Estimates next gameweek's FPL points as a sum of real scoring categories, each built from a signal already used elsewhere on this site.
Appearance
Up to 2 points for playing 60+ minutes -- FPL's flat appearance value, scaled down by starter reliability if this player isn't a nailed starter.
Goals
Predicted points from scoring, not a goal count -- it's an expected-goals rate (this season's if they've played 450+ minutes, otherwise last season's real rate) multiplied by FPL's goal value for their position (6 for GK/DEF, 5 for MID, 4 for FWD), then adjusted for how easy or hard this fixture is and how nailed-on they are to start. Divide by the goal value to see the implied expected-goals number -- e.g. 3.1 points at a forward's 4-point rate is about 0.78 expected goals, not 3.1 goals.
Assists
Same idea as Goals, but for expected assists -- FPL values every assist at a flat 3 points regardless of position, so divide by 3 to see the implied expected-assists number.
Clean sheet
Chance of a clean sheet, based purely on this fixture's difficulty rating (not this team's own recent results) -- 40% for an easy fixture, 28% for an even one, 16% for a hard one -- multiplied by FPL's clean sheet value for this position (4 points for GK/DEF, 1 for MID, 0 for FWD, since forwards don't earn clean sheet points at all).
Bonus
Expected bonus points, from this player's bonus-points rate (this season's once they've played 450+ minutes, otherwise last season's real rate), adjusted for how easy this fixture is.
Defcon
Chance of hitting FPL's real defensive-contribution threshold for this position (10 actions for defenders, 12 for midfielders/forwards -- tackles, interceptions, clearances and recoveries combined) multiplied by the 2 points it's worth. Based on their own defensive-actions rate, not adjusted for opponent.
Saves
Predicted save points -- this goalkeeper's saves-per-90 rate divided by 3 (FPL awards 1 point per 3 saves).
Starter reliability
How much of the last few gameweeks this player has actually been on the pitch, out of 90 minutes each game (benched weeks count as 0). Below 70% it's flagged as a rotation risk, and every prediction for that player is scaled down to match -- a great underlying rate built on a couple of late cameos doesn't mean much if they might not start at all.
FDR (fixture difficulty)
The Premier League's own 1 (very easy) to 5 (very hard) difficulty rating for this specific fixture. Known before a ball is kicked, so it's safe to use for predicting a match that hasn't happened yet -- unlike a stat from the same game being predicted.
Ranks candidates mainly by explosive potential (goal/assist threat), with recent form and fixture ease, plus a smaller bonus/Defcon-driven tiebreaker.
Explosive/90
Combined expected goals + expected assists per 90 minutes -- the leading signal in this ranking, since real captaincy value comes from ceiling/upside, not just who's in decent form. This season's rate once they've played 450+ minutes, otherwise last season's real rate.
Base score
Explosive potential (weighted as the dominant factor) plus recent form, all multiplied by how easy this fixture is. This is a ranking score, not a points prediction -- only useful for comparing players against each other, not read on its own.
Bonus/90
Real bonus points earned per 90 minutes, this season if they've played 450+ minutes, otherwise last season's real rate.
Defcon/90
This player's real defensive-actions rate per 90 minutes, shown against the threshold their position needs to hit for FPL's defensive-contribution points (10 for defenders, 12 for midfielders).
Bonus/Defcon rating
A 0-10 ranking signal blending bonus-points rate and defensive-contribution rate -- a small tiebreaker on top of explosive potential and form, never large enough on its own to outrank a genuine goal threat. Not a prediction of actual bonus points, which FPL caps at 3 per gameweek (6 if captained).
Starter reliability
How much of the last few gameweeks this player has actually been on the pitch, out of 90 minutes each game (benched weeks count as 0). Below 70% it's flagged as a rotation risk, and every prediction for that player is scaled down to match -- a great underlying rate built on a couple of late cameos doesn't mean much if they might not start at all.
Form
This season's real form once a player has 450+ minutes; below that, their real points-per-game from last season instead of FPL's live 'form' field, which resets to 0 at the season boundary and would otherwise treat one shock result as an established trend.
Save potential and clean sheet chance for every active Premier League goalkeeper.
Season saves
Total saves made this season so far -- a plain fact from FPL, not a prediction.
Recent saves/90
Saves per 90 minutes over this keeper's recent matches (or last season's rate, marked '(season)', if there isn't a real current sample yet).
Save potential
Low/Medium/High rating based on how threatening the next opponent's attack is (blending their season-long attack rating with actual goals they've scored in their last 5 games) relative to this keeper's own recent save rate. Forced to 'Low' regardless of the fixture if this keeper isn't a reliable starter.
Clean sheet %
Once real results exist: this team's own season clean sheet rate, adjusted up or down for how threatening this specific opponent is. Gameweek 1 only (before any real results exist): a baseline from this fixture's difficulty rating, scaled by each side's own price -- a real, always-available proxy for squad quality this early -- rather than a team's own tiny early-season sample (or literally the match being predicted), which isn't a fair pre-match signal yet.
Starter reliability
How much of the last few gameweeks this player has actually been on the pitch, out of 90 minutes each game (benched weeks count as 0). Below 70% it's flagged as a rotation risk, and every prediction for that player is scaled down to match -- a great underlying rate built on a couple of late cameos doesn't mean much if they might not start at all.
Doubtful badge
FPL has flagged this keeper's fitness as genuinely uncertain (not a hard injury) -- shown with their real chance-of-playing percentage. When a team's starting keeper is doubtful, a second keeper from the same team is shown too, since either could realistically start.
Predicts who's likely to hit FPL's real defensive-contribution points threshold in their next match.
Last match
Defensive actions (tackles, interceptions, clearances, recoveries) this player made in their most recent match, shown against the threshold their position needs (10 defenders, 12 midfielders/forwards) -- a fact, not a prediction.
Starter reliability
How much of the last few gameweeks this player has actually been on the pitch, out of 90 minutes each game (benched weeks count as 0). Below 70% it's flagged as a rotation risk, and every prediction for that player is scaled down to match -- a great underlying rate built on a couple of late cameos doesn't mean much if they might not start at all.
FDR (fixture difficulty)
The Premier League's own 1 (very easy) to 5 (very hard) difficulty rating for this specific fixture. Known before a ball is kicked, so it's safe to use for predicting a match that hasn't happened yet -- unlike a stat from the same game being predicted.
Defcon potential
Blends this player's own recent defensive-actions rate (55% weight) with how much defending their team is likely to do against this specific opponent (45% weight) -- a stronger or hotter-scoring opponent means more time without the ball, and more tackles/interceptions on offer, not fewer. Both signals are compared against this gameweek's own average before combining, then rated High/Medium/Low relative to everyone else being predicted that week.
Predicts who's set up for a strong Bonus Points System score in their next match.
Recent BPS/90
This player's Bonus Points System score per 90 minutes over recent matches (or last season's rate, marked '(season)', without a real current sample yet). BPS is FPL's own points-and-tackles-and-everything-else formula that decides who gets the +3/+2/+1 bonus each match.
Starter reliability
How much of the last few gameweeks this player has actually been on the pitch, out of 90 minutes each game (benched weeks count as 0). Below 70% it's flagged as a rotation risk, and every prediction for that player is scaled down to match -- a great underlying rate built on a couple of late cameos doesn't mean much if they might not start at all.
FDR (fixture difficulty)
The Premier League's own 1 (very easy) to 5 (very hard) difficulty rating for this specific fixture. Known before a ball is kicked, so it's safe to use for predicting a match that hasn't happened yet -- unlike a stat from the same game being predicted.
BPS potential
This player's recent BPS rate, adjusted for how weak the next opponent's defence is at this venue (a softer defence means more space, more attacking involvement, more BPS) -- rated High/Medium/Low relative to this gameweek's own average across every candidate. Anyone below 70% starter reliability is excluded outright, since bonus points need real minutes to earn.
A general squad-building/transfer-target list across all positions.
Starter reliability
How much of the last few gameweeks this player has actually been on the pitch, out of 90 minutes each game (benched weeks count as 0). Below 70% it's flagged as a rotation risk, and every prediction for that player is scaled down to match -- a great underlying rate built on a couple of late cameos doesn't mean much if they might not start at all.
FDR (fixture difficulty)
The Premier League's own 1 (very easy) to 5 (very hard) difficulty rating for this specific fixture. Known before a ball is kicked, so it's safe to use for predicting a match that hasn't happened yet -- unlike a stat from the same game being predicted.
Score
Form (this season's once reliable, otherwise last season's real points-per-game) multiplied by how easy this fixture is, multiplied by starter reliability. A ranking score for comparing candidates against each other -- not a points prediction.
Tracks the real defensive-contribution threshold live, fixture by fixture, as a gameweek plays out.
Green badge (hit)
This player has already reached their position's real defensive-contribution threshold this match (10 for defenders, 12 for midfielders/forwards) -- the 2 defensive-contribution points are secured.
Amber badge (close)
This player is within a few defensive actions of their threshold with the match still live -- shown as 'X short' so you can see how close they are to the 2 points.
Tracks the real Bonus Points System live, fixture by fixture, as a gameweek plays out.
BPS
This match's live Bonus Points System score -- FPL's own formula combining goals, assists, clean sheets, defensive actions, and more into one number per player, per match.
Bonus
The +3/+2/+1 bonus points this match's top 3 BPS scores earn. Ties share the bonus for every position they occupy: a 2-way tie for 1st both get +3, and the next player gets +1 (2nd is skipped); 3+ tied for 1st all get +3 and nobody else scores.
Tracks real saves, penalty saves, and clean sheet progress live, fixture by fixture, as a gameweek plays out.
Saves
Real shot-stops made this match, straight from FPL's live data -- each save beyond the first 3 is worth a bonus point (every 3 saves = +1), on top of the usual clean sheet and appearance points.
Pen saved
Real penalties saved this match -- worth 5 points each, on top of ordinary saves.
Clean sheet
This keeper's team hasn't conceded (yet, if the match is still live) -- worth 4 points for a goalkeeper once they've played 60+ minutes, same as for a defender.
Predicts which players are close to a real FPL price rise or fall today.
Net transfers
Transfers in minus transfers out today -- the actual number that triggers a real FPL price change. Ranked by the size of this swing, since that's what decides whether a price actually moves, not ownership percentage.
Confidence
How large today's net transfer swing is: High is 50,000+, Medium is 10,000+, Low is anything smaller. FPL's own price-change algorithm isn't public, so this is a confidence band on the swing itself, not a guaranteed call.