FPL Receipts

The starting point is FPL's own expected-points estimate for a player (or a price-based baseline where FPL hasn't published one yet). That gets scaled three ways:

• Fixture difficulty — the easiest fixtures add up to +25%, the hardest cut up to −25%, and a blank gameweek counts zero.
• A hand-researched context layer (July 2026, updated through the season) — new managers and their style fit, key signings and departures, promotion status. Each nudges a club's outlook by up to ±5% and feeds the written reasoning you see on every suggestion. A handful of player-level nudges (capped at ±10%) cover the summer role changes last season's numbers can't see — a striker becoming his new manager's focal point, set-piece duty inherited from a departed taker. Every nudged player says so, with the size of the nudge, in his "why" line — and the nudges are marked against reality by the same public model audit as everything else.
• Start chance — the odds he's actually on the pitch, from FPL's availability flags and chance-of-playing percentage, real minutes once the season is running, where his price sits in his club's pecking order at his position (FPL prices first-choice players above their understudies), and how many managers own him. This season's minutes and starts are blended in week by week — one gameweek is a hint, six is the answer — so a benching in GW1 lowers a player without erasing a full season of evidence.
• Defensive contribution (new for 2025/26 scoring) — two points for 10+ clearances, blocks, interceptions and tackles in a match (12+ for midfielders and forwards, recoveries included). The model turns each player's per-90 action rate into a probability of clearing the line every match, and prices it in — which is what separates a Gabriel or a Caicedo from a look-alike on price alone.
• Set-piece duty — FPL's own record of who is first choice on penalties, direct free-kicks and corners this season. A small nudge (capped at +7% in total) rather than a full penalty term, because last season's xG already contains the penalties a regular taker took; the nudge covers a player who has the duty now and didn't then. Always disclosed in the why-line.
• Last season's numbers, for everyone — the prior-season xG, xA, xGC and points per 90 the model anchors to are rebuilt from FPL's own player histories, so a manager arriving in October gets the same model as one who was here in July.

Price forecasts are FPL's, not ours. The "likely to rise tonight / fall tomorrow" chips on your squad come straight from the official three-day price-change projection FPL publishes for every player. We show it as it is and label it as theirs.

Start chance is why a cheap non-starter stops looking like a bargain. Every suggestion tells you whether he's nailed on, a rotation risk or a bench risk.

Minutes aren't one number. The model splits them into the chance he plays at all (appearance point), the chance he plays an hour (the second appearance point, and the clean sheet — you keep nothing from the bench after 59 minutes), and his expected minutes (which scale his attacking volume). A single "will he play" percentage multiplying everything gets all three subtly wrong, which is exactly what the leading paid models are careful about.

And we mark our own homework. Before every deadline the model's projections for your fifteen are snapshotted; once the gameweek is scored they're marked against reality, next to FPL's own free estimate and a dumb price-based guess — win or lose, on your Receipts tab. If we can't beat those two, you deserve to know.

Once real gameweeks have been played, a per-player calibration pulls the estimate towards what he has actually returned per 90 minutes — bounded, so it corrects rather than overrides.

Squads are scored the way FPL actually pays out, not by adding up all fifteen players: your best legal XI, plus the captain counted twice, plus 15% weight on the bench for autosubs and Bench Boost. That's why the builder loads money into the eleven and fills the bench with cheap enablers — adding up all fifteen would reward spending on players who mostly don't score.

The captain slot is priced on ceiling, not average. What the armband actually buys is the haul — a doubled 13 the week it lands — and hauls come from goals and assists, not from appearance points and clean sheets. So a player whose projection is mostly attacking returns is worth more as a captain option than his average suggests, which is why the builder pays premium prices for premium captains instead of treating a £15.5m striker and two mid-priced midfielders as interchangeable.

There are two build styles, and the toggle matters. Match the field (the default) breaks near-ties toward the players most managers actually hold: rank is measured against the field, and the template player's haul hurts you double if you don't own him — that risk is real even though it never shows up in a projection. Pure EV ignores ownership entirely and chases maximum projected points — the differential build, for when you're chasing rank rather than protecting it. The gap between the two builds is small in projected points and large in risk profile, and that's the honest trade.

Finding that squad uses a beam search: it carries the eight strongest squads forward at once and will accept a temporarily worse move. That's what lets it find upgrades needing two steps — selling a midfielder to afford a premium forward — which a one-swap-at-a-time builder can never see, because step one on its own looks like a loss.

Constraints are the real ones: 2 GK / 5 DEF / 5 MID / 3 FWD, max 3 per club, inside your budget.

Your fifteen are benchmarked against the best £100m draft the engine can find, across seven weighted checks:

• Starting XI strength — 30. Your best legal XI's projection over the next six gameweeks, against the benchmark's XI.
• Fixture run — 15. Average difficulty across your squad's next six.
• Value for money — 15. Projected points per £m against the benchmark's efficiency.
• Captaincy ceiling — 10. Your best armband option against the highest-projecting player in the game.
• Bench strength — 10. Kept separate rather than folded in, so punting your bench shows its real cost on autosub weeks instead of being hidden.
• Availability — 10. How many of your players carry an injury or doubt flag.
• Balance & risk — 10. Club stacking, how many clubs you cover, attacking weighting.

Three of those are ratios against the benchmark rather than fixed targets, because "good" in FPL isn't absolute — it moves with the market, and a benchmark rebuilt from live prices moves with it.

Each chip is really the same question in a different form: which single gameweek maximises one specific quantity, given your fifteen? All 38 gameweeks are scanned against your own players and fixtures:

• Bench Boost — projected points from your four bench players that week.
• Triple Captain — one player's single-gameweek ceiling.
• Free Hit — the gap between your squad's worst week and a purpose-built replacement.
• Wildcard — fixture difficulty after a swing minus difficulty before it.

Doubles and blanks don't exist at the start of a season — they appear once cup runs force postponements. The scan recomputes from live fixture data every time you open it, so the plan moves with them.

The plan rolls your real fifteen forward, applying the transfers you enter. It tracks free transfers properly: one per gameweek, banked to a maximum of five, and anything beyond what you've banked costs 4 points. The cost of a move therefore depends on moves you planned weeks earlier, which is the part that's genuinely hard to do by hand.

Each week's XI is picked on that week's projections rather than a six-week total, so a player who blanks drops out of the XI instead of walking into it. Chips are applied in line: Wildcard and Free Hit make that week's transfers free, Bench Boost adds the bench, Triple Captain adds a third copy of the armband, and Free Hit reverts your squad afterwards.

Sales inside the plan use FPL's real rule: you get back your buy price plus half the profit, rounded down — a £0.4m rise is £0.2m of spending power, not £0.4m. Not modelled, deliberately: future price rises and falls. Predicting them would imply precision the projections don't have, and the official app predicts prices for free anyway.

EO is ownership plus captaincy, because captains count twice. It's the difference between a haul that wins you rank and one that merely keeps up.

Ownership is measured — it comes straight from the FPL API. Captaincy is estimated, because FPL doesn't publish it. The estimate assumes managers overwhelmingly captain the highest-projected player they own, and distributes captaincy across the realistic candidates accordingly. Every EO figure in the app is labelled an estimate for that reason.

On your receipts, the EO shown is the one captured before that deadline and is never recalculated later — recomputing it against today's projections would answer a different question and quietly rewrite the record.

Everything stays on this device. Your squad, gameweek log, chip plan, forward plan, saved drafts and receipts are stored in this browser and nowhere else. There's no account, no email, no password, and your FPL login is never involved — importing a squad uses only your public Team ID. Nothing about your team is ever uploaded.

The trade-off is real and worth knowing: clearing your browser data deletes it. It also won't follow you to another device. Use Export data on the My Squad tab to download a backup file, and Import to restore it — the file re-matches players against live FPL data, so an August backup still works in December.

If you've bought Pro, your licence key is the durable record, not this browser. Clearing your cache doesn't lose it — re-enter the key from your Gumroad receipt and it reactivates.

Accounts and cross-device sync are coming this season, and will stay optional: the tool will always work with no login.