FPL God Mode 2026/27

A squad, and the tested reasoning behind it. Click any player's name for ranked alternatives. Every claim in here has been backtested — including the ones that failed, which are labelled.

GW1 deadline
Fri 21 Aug · 18:30 BST

players · 20 clubs · four seasons of gameweek data · database audited — 4 players the ingest reported and never added, 31 line-up sources dated 19–21 Aug, 49 start probabilities moved, 5 predicted XIs discarded for naming players who are not at the club · team news to · the between-gameweek protocol keeps it that way

Below is a complete, legal squad. If you agree with it you are done — click any name to swap someone out.
How the rest works
  1. Disagree with an assumption? Open step 2 under the squad and flip it — the Risks tab then shows what that belief is worth to you.
  2. Committed to a player? Type the name in step 3 and the optimiser builds around them.
  3. Want a season plan? Research → Gameweek plan places both chip sets and ranks your transfers.
  4. Sceptical? Method → What is tested lists every claim that survived a backtest, and the four that failed.

Start here — the squad I would pick

A complete, legal answer before you touch anything.

Search and drag in any player — all ${P.length} in the database
Your deadline checklist — what to check before Friday
What a season with this squad actually looks like
Why each of the 15 — the full reasoning
Captain Haaland. And it matters far less than you have been told.

Vice goes to .

But the honest number is this: I played 143 squads through all 38 gameweeks of real data and varied one lever at a time. Changing the squad moved the season by 280 points of standard deviation. Changing the captaincy policy, holding the squad fixed, moved it by 41 — perfect hindsight captaincy scored 1,178, a sensible policy 1,040, and the worst defensible policy 966. Squad choice matters roughly eight times more than captaincy.

That inverts the brief this tool was built from, which ranked captaincy first. Weekly variance is not season variance: the armband feels decisive every Saturday and is nearly noise across a year. Captain the best player you own, do not agonise, and spend the saved attention on the fifteen.

Why this fifteen — the live calls, the Community Shield result, and what live re-verification found as at ${(D.verified&&D.verified.at?D.verified.at.slice(0,10):'')}
Read this as a dated record, not as a description of the squad above.

The calls below were made on 16 August against the fifteen as it stood that day, which included Semenyo, Igor Thiago and Virgil. The squad has since been re-optimised on the corrected engine — four-season points per start, the promoted-club concession prior, measured newcomer rates and the unverified-start discount — and none of those three survived it. The current fifteen is the one on the pitch above; the per-player reasoning for it is in Why each of the 15.

The paired-season measurements are kept because they are still true of the comparison they were run on: 20,000 identical simulated gameweeks, only the swapped player differing. Deleting a measurement because the squad moved on would be the marketing-page version of a corrections log. What has changed is which squad the question was being asked about.

Three live calls: is a third Manchester City asset worth it?

You have Haaland and Semenyo, so the club limit allows exactly one more. I tested all three candidates over 20,000 paired seasons — identical simulated gameweeks, same team-week shocks, so only the swapped player differs.

SwapMeanWinsVerdict
O'Reilly for Virgil−1633.4% Your call, and a real one. It breaks even the moment he is nailed: at a 97% appearance rate and a 120-point projection the swap turns positive. Last season was 24 starts and 4 absence spells against Virgil's 33 and zero.
Gvardiol for Virgil−429.5% No. Highest xGI of any City defender (15.7) — and 14 starts with 6 absence spells. Brilliance you do not get to field.
Foden for Szoboszlai−564.9% No. 67 goal involvements and an 18.2% haul rate, the best ceiling here after Haaland — but 19 starts against Szoboszlai's 31.

All three fail for the same reason, and it is the one thing this tool has actually proved: minutes. Each is a higher-ceiling footballer than the man he replaces and each plays less. That is also why O'Reilly is the live one — he is the closest to nailed, and the Community Shield is evidence his role has grown that last season's numbers cannot see.

Note what did not save them: the tail. O'Reilly's 99th percentile is 2,128 against 2,134 — the ceiling argument does not survive contact with the simulation, because a player who misses eight games cannot haul in them.

Arsenal 3–0 Manchester City, 16 August — and Calafiori scored after one minute.

All three open questions in this squad resolved in its favour. Haaland started — every pre-match predicted XI had him benched and my own note repeated it, so that was my error, not the consensus's. Calafiori started at left-back with Hincapié on the bench, and scored. Semenyo started for City. Havertz scored Arsenal's second having been left out of both major predicted XIs.

Caveat: Arteta rotated hard — Saka, Rice, Zubimendi, Merino, Eze and Gyökeres all began on the bench, and Rice and Saka have only just returned from the World Cup. A Shield XI proves fitness, not GW1 selection. City's side was much closer to first choice than Arsenal's.

What a live re-verification pass found, and it was not nothing.

I re-checked every squad member's club against primary sources rather than the database — the discipline I wrote after being caught out by Bruno Guimarães. It found Nørgaard listed at Arsenal when he signed for Everton on 5 August, and he has now replaced Röhl in this squad on merit.

Pulling predicted XIs from three sources instead of one changed more: Röhl is called a midfielder by one, an attacker by another and a winger with no guaranteed start by a third. Gyökeres was in both major predicted XIs and was benched. Liverpool's penalty order is genuinely disputed — one specialist source ranks Isak first, another ranks Szoboszlai first and calls the hierarchy contested; no public Iraola statement exists. Consensus was already priced; the disagreement is where the information was.

Transfer risk still live until the window shuts on 1 September: Man Utd enquired about Gibbs-White (8 Aug); Chelsea and Man Utd about Thiago (7 Aug, £80m ask); and Barcelona's funding blocker for João Pedro cleared when Ferran Torres joined PSG on 15 August. None has moved. All three are in this squad.

1  One-click builds

Each preset sets the vector and the sliders, then optimises. Anything you have locked survives every preset.

2  What do you believe will happen?

Sixteen contested questions from the research. Flipping one moves only the projections it touches. Defaults are what the evidence currently supports — the Risks tab shows what each belief is worth to your squad.

3  Bet on specific players

Type anyone you are committed to — Isak, Wirtz, Tonali — and the optimiser builds the best legal squad around them. They stay through every preset and every scenario change.

4  Tune the optimiser rarely needed

Most of this you should leave alone. Three of these settings encode strategies that were tested and failed. The defaults are the ones that survived: Edge vector, no ownership cap, low ceiling weighting, minutes weighting on. Change them to reproduce the tests, not to improve the squad.

Which vector are you maximising?

The same ${P.length} players, scored seven ways. Two of the seven are kept only so you can see what they cost.

Sliders

Tested and it does not work. Capping ownership lowered the mean, lowered the tail, and lowered the margin over the template in every arm. At 120% not one squad in 4,249 reached the top 1%. Left open by default.
Also tested, also fails. Filtering on prior-season haul rate cost 95 points at a 10% bar and 436 at 20%, and cut the 99th percentile too. Prior haul rate does not persist. Kept low by default.
Spend less and the rest is banked. Team value barely matters — 270 players fell in price last season against 55 that rose.
How far the gameweek plan projects.
Discounts anyone outside their club's predicted GW1 XI — 50/50 calls ×0.72, non-starters ×0.42. Predicted minutes beat every other input.

Price changes, team value, and the fixture events nobody can see yet

Three things this tool used to list as unmodelled. Two of them are now measured, and the measurements say something different from the folklore. The third is a discipline, not a model.

Gameweek plan

Built from your scenario, your squad and the real fixture list for all 38 gameweeks. Both chip sets are placed below; the gameweek-by-gameweek rows follow your Horizon slider in Build, which now runs to GW38.

Go to Build and press Build my squad.

Shortlist

Every player, scored under your current scenario. Score is the scenario-adjusted projection; Value is score per £m; Haul% is the best full season in four years — the ceiling measure that matters for a title, not a rank.

PlayerClubPos£Own% GW1–12 SeasonVal RoleEdgeHaul%SpellsBlank% GW1-8 fixtures

Spells = most absence runs in any season. 3+ is the chronic flag; minutes load does not predict injury, prior absence does.

−17
Bonus points defenders lose to a rule change nobody has priced
FPL cut the reward for clearances, blocks and interceptions from one BPS per two actions to one per three for 2026/27. I replayed all 380 matches of last season under the new rule. Every projection built on raw 2025-26 points overrates centre-backs and underrates goalkeepers.

The vault — data I computed because nobody publishes it

Two things elite managers want and no public tool provides: last season restated under the new bonus rules, and defensive contributions modelled as a probability of clearing the threshold rather than a season total. Both are computed here from 380 matches of raw gameweek data.

20
Clubs, four seasons each, sorted by opening fixtures
Click a club to open its full analysis — squad, risks, who to target and when. The second-half split shows who was actually trending, which the season table hides.

All 20 clubs

Four-season record, the second-half split that shows who was actually trending, GW1–10 fixtures, and the manager situation. Sorted by GW1–8 fixture difficulty.

Club points model — how your scenario moves each club

Expected FPL points from each club's best assets under the switches you have set: top keeper, top three defenders, top three midfielders, top two forwards. Toggle a belief on the Build tab and watch a club move. Click any column to re-sort.

ClubTotalGKDefence MidfieldAttackBest asset Best haul%GW1-8 FDRTemplate load

Template load = total ownership of that club's top six assets. High numbers are where the crowd is; that is where you gain nothing by matching and lose heavily by missing.

Club by club

The three promoted clubs — I tried to tell them apart, and could not

4
Computed screens, not hunches
Each screen below runs over four seasons of data and names the way it fails. Low ownership alone wins nothing — every screen also requires a confirmed starting role.

Blue ocean — who might beat their projection

Projections are backward-looking by construction. These four screens look for the players most likely to outperform them, and every one is a computed screen over four seasons of data — not a hunch. Each has a stated failure mode, because each is a bet.

Fixture difficulty, by position — not FPL's

FPL publishes one difficulty number per fixture and gives it to a keeper and a striker alike. That is the wrong shape: a trip to a side that concedes chances but rarely scores is easy for a forward and easy for a defender at the same time, and one number cannot say both. This grid is built from the two models the projection already runs on — the clean-sheet model for defence, the measured opponent multiplier for attack — so it can never disagree with the pitch.

How to read a cell. The tier is the opponent alone, banded on the distribution of all 240 fixtures in the window rather than on a hand-drawn scale: tier 1 is the easiest fifth, tier 5 the hardest. The number beneath it is what that fixture is worth in points to a player of the selected position, priced through FPL's rules. Difficulty is shared between GK and DEF (both ask "how much does this opponent score?") and between MID and FWD (both ask "how much do they concede?"). The points differ for all four, because the rules do.

The value vector — points per million

A rank competition is played under a budget, so the axis a squad is actually built on is not projected points but projected points per pound. Ranked below over the same window, with the raw projection beside it so you can see which players are cheap because they are good value and which are cheap because they are bad.

Trendsetters — what the top of the world actually owns

The top of the Overall league, read straight from FPL after the deadline, against live field ownership from FPL's own selected_by_percent. The gap is the point: rank moves against effective ownership, so a player the leaders own and the field does not is where a rank is made or lost. One scan is a photograph. From the second gameweek onward the move column is the trend — who the leaders bought this week.

This costs real API calls. One standings call plus one squad call per manager scanned, four at a time, cached for the gameweek. That is why it is a button.

Defence, rated against the opponent

The clean-sheet model this build already carries, applied to one fixture instead of averaged over eight. λ = base × att[opponent] × dfn[own club], where base is the league's home or away scoring mean, and P(clean sheet) = e−λ. Since patch268 the projection reads this, through the component engine below: the clean sheet is one of its nine terms. Until then expected points came from a four-season points-per-start rate with no clean-sheet term at all, and this panel was a second opinion nothing consumed.

Points, by FPL rule

A second engine, built bottom-up from FPL's own 2026/27 scoring table — nine terms, each one a line of the rules, each computed from a fixture-specific probability rather than from a historical average. Goals conceded and saves are summed over their Poisson distributions rather than divided, because FPL floors them: −1 per two conceded, +1 per three saves. Appearance, the clean sheet and the defensive contribution all hang off the same 60-minute event.
Validated against the four seasons it is built from: on players with 7,000+ minutes of history the median residual is +0.03 and the mean absolute error 0.28. Since patch268 this IS the projection — or rather, a measured share of it. Every number on this page is now w × rules + (1 − w) × points-per-start, where w = minutes / (minutes + 1800) is how much Premier League evidence stands behind the rules-based estimate for that player. A 7,000-minute player is 80% rules; the median 2,600-minute player 59%; a promoted-club player reached through his Championship record is capped at 45%, because a translated rate is a prior and not a measurement. The weight is the sample size because that is exactly what this engine's measured error depends on: its residual runs −0.81 at 450–1,200 minutes and +0.03 at 7,000+.
What it bought, measured. On 114 nailed starters with 3,000+ minutes, predicted against observed points per 90: the old projection ran at 0.752 of observed — it under-rated a good player by a quarter. The blend runs at 0.895. Some of the remaining gap is selection (nailed starters with long records are better than average) and it has deliberately not been closed with a fudge factor.
Add ?engine=1 to the address to put the old projection back and compare.

38
Gameweeks, every club, coloured by difficulty
Sorted by the window you choose. Use this for planning transfers, not for reweighting projections — adding fixture difficulty to a single-gameweek estimate made the model worse, 1.87 to 1.92 mean absolute error.

Fixture ticker — all 38 gameweeks

Every club, every fixture, coloured by difficulty. Sorted by the window you choose. Click a club row to jump to its full analysis; click a player to bring him into your squad.

Rotation planner

The cheap-defence trick, generalised and scored in points. Choose how many clubs you own and how many of them start each week; every combination is evaluated, and in each gameweek the highest-expected player in the pool is the one fielded. Other rotation tools rank on a 1-to-5 fixture rating. This ranks on expected points — the player's own scoring rate, his measured chance of not appearing, and a measured opponent factor including a prior for the three promoted clubs that have no Premier League record.

Who to target in this window

The best assets at the clubs with the best run over the gameweeks you selected, weighted by how much each opponent concedes to that position.

+42%
What opponent strength is actually worth
It makes single-gameweek point estimates worse — but it ranks players better, and ranking is what a transfer is. Tested on GW20–38 with everything fitted on GW1–19 only: rank correlation 0.089 → 0.126, better in 13 of 19 gameweeks.

Gameweek targets — who to own, and when

For each gameweek: the softest defences to attack, and the players best placed to punish them. Concession factors are measured over a full season — a club at 1.15 gave up 15% more than an average opponent to that position.

8
Chips, in two sets, and the first four expire at GW19
An unplayed chip scores zero, and that is the most common self-inflicted loss in FPL. Set a reminder at GW15.

Chip strategy — both sets, all 38 gameweeks

Eight chips in two sets. The first set expires at the GW19 deadline — an unplayed chip scores zero, and that is the most common self-inflicted wound in FPL. Well-timed chips are worth roughly 90 points across a season; captaincy alone is worth more, so plan them once and then stop thinking about them.

The honest caveat, up front The published fixture list has no blank or double gameweeks in it — I checked all 380 fixtures. Doubles and blanks only emerge once cup rounds are drawn and games are postponed, typically from December. Any plan that claims to know your Free Hit and Bench Boost weeks today is guessing. What follows anchors on real fixture runs, which are knowable, and tells you when to revisit the rest.
Percentage of the tool's own numbers that rest on more than one season
Every projection you have ever been shown by anyone rests on a choice about how far back to look. Here is mine, audited, including where it is thin.

What the data actually is

A full inventory: what is fresh, what is four seasons deep, what is one season and pretending to be more, and what is simply missing. This page exists because a tool that will not say where it is weak cannot be trusted where it is strong.

Players in the game with no Premier League record at all
Nearly a third of the database. For every one of them the projection is a price-band average and nothing else. This is the page where I stop pretending otherwise and go and look.

The newcomers

A human cannot check 177 players against pre-season team sheets, transfer fees, squad numbers and foreign-league output in the week before a deadline. That is the one part of this job where being a machine is a straightforward advantage. Here is what the check found.

What a start is actually worth, by where he came from

4%
Of £4.0m players who reach 1,800 minutes in a season
Everything written about "value per million" is measured on the ones who played. That is the wrong denominator, and it is why cheap enablers look like free money and mostly are not.

Where the money goes

The £100m question is not which players are good. It is what a pound buys. I measured that two ways on four seasons of data, and the two answers point in opposite directions — which is the whole point.

Points a season the bench contributes through autosubs
I went looking for this number in the published literature and it does not exist. Every guide explains how autosubs work; none of them says what they are worth. So I measured it.

The other four

Fifteen players, eleven of whom score. The four who do not are the least-analysed part of the game, and the place most managers quietly leak points.

Points it cost to spend a transfer purely so it would not be wasted
The five-transfer cap creates a trap: you feel you must use one before it lapses. Letting it lapse was worth more.

Free transfers, banking, and the cap at five

You get one free transfer a gameweek and can bank up to five. Beyond five, the next one is simply lost. That single rule generates most of the weekly agonising in this game, and it is testable.

0.15
R² of the best public expected-points model, at gameweek level
Against a theoretical perfect model at 0.17. Almost everything separating the good tools from the great ones has already been squeezed out. What is left is discipline.

Playing the field

Ownership, the template, the armband, and the uncomfortable truth about how much any of the modelling is really worth. This page is the strategy literature, checked, sourced, and where possible tested.

The armband: I tested the differential, and the differential loses

Expected points over gameweeks 1 to 12, holding your fifteen
Every week costed: the eleven that starts, the armband, the bench, the best transfer on the board and what each chip is actually worth. And the one number the other tools will not print — how much of it is real.

The twelve-week plan

This is a calculator, not a prophecy. It scores your exact fifteen against the published fixtures for GW1–12, picks the highest legal eleven and the armband each week, and prices every decision available to you. Then it applies the discount that a walk-forward test says the whole exercise deserves.

GW12 the first-set chips expire at GW19

Week by week

Why each week looks like this

The table above prices every week. This says why — the fixture behind the armband, the margin it won by, the slot that is closest to changing, and what would overturn it before the deadline.

What the transfer machine is worth — measured, not asserted

The chips

Transfer search — the whole squad, not one slot

Every package of up to three moves, including the ones where selling somebody expensive funds two upgrades elsewhere. Each is scored on the whole team's expected points with the eleven re-picked every gameweek and the armband applied — not on the two players swapped. Short-window fixture plays show up because every package is priced over one week, three and the full horizon at once.

The transfer ledger — a decision every week, not a list of options

The plan

Known problems with this tool that are still open
Not a roadmap. A list of the places where this site is wrong, unvalidated, or running on a prior instead of evidence — ordered by how many points are at stake, with the fixed ones kept visible so the corrections are auditable rather than quietly disappeared.

What is still broken

Every tool like this has a gap between what it measures well and what it asserts confidently. The difference is whether the gap is written down. Each item below says what is wrong, what the evidence is, and what fixing it requires — and where I have measured the cost in points, it is stated; where I have not, it says so instead of guessing.

Share of a whole season the median player scores inside his best six weeks
Every big score in this game arrives in a lump. The question is not whether purple patches are real — they are, and they are enormous. It is whether you can be holding the player when one starts, and that is a different question with a much less flattering answer.

Purple patches, and the plan to buy one

The instinct is sound: bank a transfer, sell elsewhere, and bring a big scorer in for the run where his fixtures open up. I have tested it on four seasons of gameweek-by-gameweek data. 1,327 player-seasons supply the runs themselves; the 573 with a complete set of rolling six-week windows supply the test of whether you can see one coming, each window scored for what its opponents actually conceded to that position that season. The patches are real. The timing is not buyable.

It works least well on exactly the players you want it for

The cost of waiting

Liverpool, and the gap in this squad

Where every club's easy run sits

The twenty biggest runs of the last four seasons

Points your squad is expected to lose to players who do not appear
Not injuries you can see. The ones the fixture list cannot tell you about — rotation, a knock on the Thursday, a manager protecting someone for Europe. Every squad leaks here and almost nobody measures it.

Minutes risk and autosubs

A projection is a number multiplied by the chance he plays. Most tools stop at the first half. This page does the second: it samples who blanks, applies FPL's real autosub rules — keeper for keeper, bench in order, the shape must stay legal — and reports what it costs you.

Build or load a squad and this page fills in.

The projection grid

Every player in your fifteen, every gameweek in the window. Read down a column for a bad week coming; read across a row for a player whose fixtures do the work for him.

Your bench order, priced

FPL substitutes in the order you set, and most managers set it by price or by gut. The first bench player who both turned out and keeps the shape legal comes on — so a forward at number one can be blocked by the three-forward cap while a defender behind him would have come on. That makes the order a small optimisation problem, not a sort. All six permutations are priced below.

How the blank probability is fitted

Fitted on GW1–19 of 2025-26, scored on GW20–38, which the fit never saw. Brier score against for assuming every player carries the league-average blank rate — a skill score of on player-gameweeks. A recency half-life of 8 gameweeks and a shrinkage of 2 pseudo-observations were chosen on a held-out block (GW20–28) and confirmed on GW29–38; Platt recalibration was tried on top and made it worse, so it is not applied.
A season blank rate is the wrong number for a player who is fit today. It includes the eight weeks he spent injured in March. What you want is the chance he misses next gameweek given he started the last one — measured on 3,407 such cases in 2025-26 at 9.3%, and still strongly graded: 6% for the reliable, 26% for the chronically absent. The fitted relationship is P(blank | started) = 0.151 × season rate + 0.050, and that is the number this page uses for every player you have marked as a predicted starter. Players marked 50/50 get the average of the two; players not expected to start keep their raw season rate, because for them the season rate is the story.
Players with no Premier League record carry a measured prior, not a guess. I took every player who started four of his club's first six gameweeks in 2024-25 and 2025-26 and split them by where they came from: established starters blanked 13.2% of gameweeks, new signings to existing clubs 14.5%, promoted-club players with no PL record 14.9%. So a predicted starter with no record is given 15%. Those players are marked prior in the table.
Checked against reality, and it still runs about 40% high. I took the eleven most-owned players in the game in each of the 38 gameweeks of 2025-26 and counted how many recorded no minutes: 0.66 of 11 per gameweek. This model, applied to the same eleven, expects 0.92. The gap is not noise — it is that the most-owned players are, by construction, the ones managers have already transferred out when they got injured. So read the number below as the cost of holding this squad and never reacting to team news. Reacting is what closes the gap, and it is free.
What this does not model. Points conditional on playing are taken from last season's points-per-start, so a player who comes on for twenty minutes is treated as having started. That biases the cost of a blank downwards. Blanks are also drawn independently; in reality a club's rotation is correlated, which widens the tail. And no pre-season model knows who is carrying a knock — the first team sheet of the season will be worth more than any of this.
15
Players each, one season, one honest scoreboard
The shared players cancel out all season. Only the differences decide it — and the gameweek log below is there to show which kind of judgement won.

AI vs you — two squads, one season

My fifteen against yours, tracked from gameweek to gameweek. The point is not to find out who wins. It is to find out which kind of judgement wins, on which sort of decision — because in the four disagreements settled so far — all of them last season — you were right and I was wrong, and every one was about minutes. Nothing has been played in 2026-27 yet, so this season’s record is 0–0.

The registered scoreboard

God Mode plays the season as a real FPL entry, with a deadline it cannot miss and a squad it cannot revise afterwards. That is the only version of this comparison worth anything: an optimiser that re-solves on every page load never has to commit, and so can never lose. Both rows below are read from FPL's own history — nothing here is typed in.

Gameweek log

Enter both scores after each deadline. FPL gives you yours; mine is the same fifteen scored the same way. The running gap is the only honest scoreboard either of us has.

0.367
Correlation between team-mates in the same gameweek
Measured over 8,564 player-gameweeks. Every other FPL optimiser, including mine until recently, treats players as independent. They are not.

Synergy — what your players do to each other

Every optimiser in FPL, including mine until now, treats players as independent. They are not. A player's score correlates r = 0.367 with his teammates' that same week — measured across 8,564 player-gameweeks. When a team wins 4–0, the striker, the creator and the defender all haul together.

Why this matters more than it sounds Independent draws understate the right tail of a stacked squad. The tool has been warning you about club concentration as a risk and never once crediting the upside. Both are true: a stack blanks together and hauls together. For a manager maximising expected points that is a wash. For a manager who needs one 120-point week, it is the entire point.

Build or load a squad to see its correlation structure.

Risk register

Two kinds of risk: the ones baked into any season, and the ones specific to the squad you just built. Build a squad and this page audits it.

Build a squad first and the sensitivity analysis appears here.

The structural risks — true whatever you pick

ScenarioLikelihoodDamageResponse
All differentials miss~30%200–400 ptsDetect at GW6 on role, not returns. Revert to template. Never add more differentials to recover — that is how a bad season becomes a catastrophic one.
The dominant asset fails or is injuredModerateSharedDo nothing immediately. 75% of the field owns him, so it is shared risk. Panic-selling converts shared risk into unique risk.
Outside 500k by GW10CommonRecoverableReassess strategy, not personnel. Max one transfer per gameweek. Every worst case is made worse by activity.
The regime-change thesis is wrongReal — nine new managers100–200 ptsRevert to continuity clubs: Everton, Brentford, Leeds, Arsenal, Villa, Brighton, Sunderland.
Multiple simultaneous injuriesLikely onceModerateThe one case where a −8 is justified. Never field fewer than eleven.
Chips stranded at GW19Entirely avoidable~45 ptsPlay all four first-set chips by GW17. Early disruption is real but small and none of it is chip-shaped: 2024-25 blanked at GW15, three of four seasons had one, and the official 2026-27 fixture list has no blank and no double anywhere in GW1–12. Set a reminder at GW15.
Transfer churn under pressureHistorically OctoberCompoundingBudget 4–6 hits all season. Above six by December means you are chasing. The 2.3m-rank season required being short the leader and churning.
The GW7 rotation trapCertain1–2 blanksAll nine European clubs play two days after a European night. Plan and execute the transfer in GW6.

Data risks in this tool

All fifteen clubs verified against live sources, 16 August 2026. Every player in the recommended squad was checked against the live FPL API team list and at least one independent source — club official sites, ESPN, Sky Sports, the Premier League squad pages. Nothing had moved. The four transfers already baked into this database all confirmed: Kelleher to Brentford, Dubravka to Spurs (from Burnley), Semenyo to Man City (£64m, January), Tonali to Spurs (club-record £100m). Gibbs-White signed a three-year extension at Forest to 2028, which ended the Spurs release-clause dispute. Calafiori confirmed at Arsenal by the live API itself.
The player data can still be wrong about clubs, and the window is open until 1 September. Bruno Guimarães signed for Arsenal on 9 August for £75m and both the public archive and the live FPL API still had him at Newcastle six days later. That is the failure mode: the archive lags, and a verified squad on Sunday is not a verified squad on Friday. Gibbs-White is the one still attracting bids — Aston Villa, Man Utd and Chelsea all linked in the first week of August, contract extension notwithstanding. Re-check before the deadline.
The promoted-club blind spot is fixed. Coventry, Hull and Ipswich have no Premier League record, and this tool used to score them as exactly average opponents — which meant it could not see its own best fixtures. Measured across the three promotion cohorts of 2023-24, 2024-25 and 2025-26 (3,549 player-gameweeks), newly promoted clubs concede ×1.31 to defenders, ×1.21 to midfielders, ×1.16 to forwards and ×1.12 to keepers. That prior is now applied everywhere, and it moved the captaincy answer in four separate gameweeks.
The opponent table still describes last season's league. Three of 2025-26's softest opponents were relegated, and the promoted prior above is a league-wide average, not a read on these three clubs specifically. Rebuild it at GW6 on this season's results.
Every start estimate is judgement, not measurement. So is every scenario multiplier. The projections underneath them are computed; the layer on top is a considered guess, and it is the layer most likely to be wrong.
The gap has closed, and this note is kept rather than deleted. It used to read "twenty players are invisible: the live API carries 587 elements, this database has 567". The database now carries ${P.length}, and the late signings it named — Tzolakis, Mendy, Gonzalo García, Diomande — are in the optimiser. What is still true is the shape of the risk: a database built from a snapshot goes stale inside a transfer window, and the way you find out is by counting rather than by assuming.
The optimiser is a local search, not a proof. It restarts five times and hill-climbs. It will find a very good squad; it will not always find the best one, and re-running can produce slightly different answers.
89%
First live test — Community Shield team sheets, 16 August
32 of 36 predicted-XI calls correct. Six claims below passed a backtest, four failed — and three of those four had already shipped in this tool. The failures are listed with the same prominence as the wins.

What is tested — the full ledger

Every claim this tool acts on, the test it was put through, the number that came back, and whether it survived. Six passed. Four failed, and three of those four had already shipped. The failures are listed with the same prominence as the successes, because a tool that only reports its wins is a tool you cannot calibrate.

How the testing works, in one paragraph.

Every test is walk-forward: a model is fitted on seasons before the one it is scored against, never on the season it is predicting. Where a correction is fitted (the position debias, the price blend weight), the correction itself is fitted out of fold — on data the scoring season never touched. That second control is what caught the debias failure; an in-sample fit made it look like an improvement. Squad-level claims are tested by building thousands of legal squads under a constraint, playing them through all 38 gameweeks of the real 2025-26 gameweek data, and comparing distributions — not just means, because a strategy that wins on mean and loses in the top percentile is useless to someone trying to win.

Passed — and applied

ClaimTestResult
Predicted minutes beat everything else you can add to a gameweek model Per-gameweek model, with and without a predicted-XI weight, walk-forward PASS. Correlation 0.232 → 0.245, mean absolute error 2.31 → 1.87 — a 19% reduction in gameweek error. The single best-validated finding here, and it came from your correction on Hincapié, not from me. On by default; the toggle is in Build.
FPL's own price carries information the model does not Blend model projection with price×15 at varying weights, scored across three walk-forward folds PASS, replicated. The model alone ran r ≈ 0.45–0.56. Price×15 alone beat the model on correlation in two of three folds. A 40% model / 60% price blend beat both, in both folds where it could be tested. Every projection in this tool is that blend; the gap between the two is what the Edge vector ranks by.
Players with no Premier League history need their own line 542 no-history players, fitted out of sample PASS. They average 36 points and 5% clear 120. Price×15 was worse than a constant for this group. Fitted line: 26 × price − 86. Every cold-start player in here is flagged, because that line is a population average and says nothing about any individual.
Squad choice dominates captaincy 143 optimiser-built squads × 38 real gameweeks, one lever varied at a time PASS, and it inverts the brief. Squad choice: SD 280, spread 1,126, top decile minus median 329. Captaincy policy with the squad held fixed: mean spread 41, maximum 72. Roughly 8:1. Weekly variance is not season variance.
Spreading beats stacking Squads capped at 1 per club vs 3 per club, full-season distribution PASS for spreading. Mean 1,503 vs 1,483; 99th percentile 1,776 vs 1,724. Stacking loses the tail as well as the mean — the opposite of the usual argument for it. The 3-per-club FPL rule is not a constraint you should be trying to max out.
Pick keepers on clean sheets, never on saves Correlation of keeper points against saves per 90 and clean-sheet rate PASS. Saves per 90 correlate −0.43 with keeper points — a save-heavy keeper is a keeper behind a bad defence. Clean sheets correlate 0.83.
First live test, 16 August 2026 — Community Shield team sheets.

The predicted-minutes layer is the tool's best-validated claim, so the first real team sheets of the season are a genuine out-of-sample test of it. Against the Arsenal and Manchester City XIs and benches: 32 of 36 hard calls correct — 89%. Three further players were flagged 50/50 and score neither way.

MissCallWhat happenedOwned
O'ReillybenchStarted at left-back for City. The costly one.22.9%
GyökeresstartBenched — but so were Saka, Rice, Zubimendi, Merino and Eze. Rotation, not demotion.13.0%
Matheus NunesstartBenched.5.0%
Rúben DiasbenchStarted and captained City. A first-choice centre-back my model had outside the XI.1.4%

What I am not going to do with this. Rúben Dias is now correctly flagged as a starter at £5.5m and 1.4% ownership, which looks like exactly the differential this tool is supposed to find. He still does not make the squad, and he should not: his projection is 87 against 109 for Calafiori at the same price, because he returns clean sheets and almost nothing else, and City's clean sheets are already priced into cheaper defenders. Low ownership is a consequence of being right about a player — it is not itself a reason. That was the finding that killed the ownership cap, and it applies to picks I like as much as to ones I do not.

One test is one test. 36 calls is a small sample, both clubs rotated for a trophy that is not the league, and the honest read is that this is consistent with the 19% error reduction rather than a confirmation of it.

Failed — three of these had already shipped in this tool

ClaimTestResult
Correcting the model's per-position bias improves it Bias fitted out of fold, then applied to a held-out season FAIL, and it had shipped. Mean absolute error got worse: 34.6 → 39.4 in one fold, 36.8 → 37.3 in the other. The forward bias flipped sign between folds, +28.0 to −13.1 — meaning it was noise being fitted, not a bias being corrected. In-sample it looked like a clear win. Removed.
Capping ownership makes you more likely to win 4,249 legal squads built under caps from 120% to 300%, all played through 38 real gameweeks FAIL, decisively, and it had shipped. At a 120% cap, zero of 4,249 squads reached the top 1%. The 300% arm won on mean (1,488), won on the tail, and won on margin over the template squad (+231 against +117). Low ownership is a consequence of being right about a player, not a method for becoming right. Cap removed; the slider is open by default.
Filtering for high haul rate raises your ceiling Minimum haul-rate bars applied to squad construction, scored out of sample FAIL, and it had shipped as a vector. A 10% bar cost 95 points; a 20% bar cost 436. It also lowered the 99th percentile — it did not even buy the ceiling it was sold on. Past haul rate is mostly past minutes and past penalties, both of which are already in the projection. The Ceiling vector is kept, labelled, so you can reproduce the failure yourself.
Adding opponent strength and home advantage sharpens the gameweek model Layered onto the minutes-weighted per-gameweek model FAIL. Mean absolute error went from 1.87 to 1.92 — worse. Fixture difficulty is real over a run of games and is already priced into who you own; adding it to a single-gameweek point estimate subtracts accuracy. Not applied. The fixture ticker is for planning transfers, not for reweighting projections.

Not tested — treat these as opinion

Three things in this tool have not survived a backtest, because I could not construct an honest one. You should discount them accordingly:

  • The 16 scenario switches. Every multiplier attached to "Spurs fix their defence" or "Iraola unlocks Wirtz" is my judgement expressed as a number. They are calibrated to be modest for that reason, but they are not evidence.
  • The conviction slider (±11% per step). The step size is a guess at how much a human read should be allowed to move a projection. It is deliberately small enough that being wrong costs you little and being right gains you little.
  • The team-week correlation factor in the simulation. The correlation itself is measured (r = 0.367 between team-mates' gameweek scores). Whether feeding it into the simulation changes any decision for the better is untested.

Everything on this page that is labelled data comes from four seasons of real gameweek results. Everything labelled judgement is mine, and you win arguments about football — so where they conflict, back yourself.

8:1
Squad choice against captaincy
Changing the squad moves a season by 329 points between median and top decile. Changing the captaincy policy, squad fixed, moves it by 41. Spend your attention on the fifteen.

Where points actually come from — tested

The operating brief this tool was built from ranked captaincy first and squad structure third. I tested it: 143 squads, each played through all 38 gameweeks of 2025–26 with real per-gameweek scores, varying one thing at a time.

LeverMeasured valueWhat it is
Squad selection329 ptsMedian squad to top-decile squad. Standard deviation 280; best minus worst 1,126. Decided entirely before a ball is kicked.
Captaincy — the ceiling139 ptsPerfect weekly hindsight against a sensible fixed policy. Nobody achieves this.
Captaincy — what is actually available41 ptsThe spread between sensible policies: best prior player, most expensive, recent form, random of your top three. All four landed within 41 points of each other over a full season.
Squad selection outweighs captaincy by roughly eight to one. This inverts the brief. Captaincy is where almost all FPL attention goes — the Sunday-morning agonising, the last-minute switch — and between reasonable choices it is worth 41 points across an entire season, a little over one point a week. Meanwhile the gap between a median squad and a top-decile squad is 329 points, and it is fixed the moment you press save.
The practical instruction Captain the obvious player and stop thinking about it. Put the recovered attention into who is in the squad and whether they are starting. If you are agonising over the armband, you are optimising a 41-point lever while a 329-point one sits unattended.

And this resolves the variance puzzle

There is no variance to buy. There is only quality to accumulate. Three separate attempts to manufacture season variance — capping ownership, filtering on haul rate, stacking clubs — all failed, and this is why. A season is 38 gameweeks, which is enough draws to average away nearly all week-to-week randomness. What survives is a 1,126-point spread that is not variance at all — it is skill, already determined at selection.

So the dispersion you need in order to win is already there. It comes from the rest of the field selecting worse than you, not from you taking bigger risks. Adding noise to a decision that rewards accuracy can only hurt.
Which makes the whole strategy one line Win by not being wrong, not by being different. Get the minutes right, take a mispricing when you find one, captain the obvious player, do not churn, do not lose a chip. Every clever alternative tested worse than that.

The individual signals, and what testing did to each

SignalStatusDetail
Predicted minutesValidatedCuts gameweek error 19%. The single largest measured effect anywhere in this tool.
Market mispricing (Edge)SurvivedThe only differentiation instrument that passed. Currently points at defenders being unpriced for defensive contributions.
Squad selection qualityDominant329 points, median to top decile.
CaptaincyReal but small41 points between sensible policies. Do not overthink it.
ChipsUntested hereRoughly 90 points by reputation. Unplayed they are worth zero, which is the only part that is certain.
Opponent difficultyFailedAdding it to the gameweek model made the error worse. Use fixtures to time a transfer, never to predict a score.
Home advantageFailedSame test, same result.
Ownership capsFailedLower mean, lower tail, smaller margin over the template.
Haul-rate filteringFailed−95 points at a 10% bar out of sample. Prior haul rate does not persist.
Club stackingFailedSpreading beat stacking on mean, 90th, 99th and best outcome.
Career-peak chasingFailedOver-predicts by 83 points. 73% of those players stay down.

Every "failed" row is a strategy this tool shipped with at some point during its construction, and every one was removed after it lost a test rather than before.

Analysis

Three readings of the fifteen on the pitch, each with the button that acts on it. A finding you cannot act on is a paragraph, and this project has written enough of those.

Triggers firing on your fifteen right now
A squad is a snapshot of what was known on the day it was picked. This is the cycle that stops it becoming one, and the log of every time it changed the plan.

Between gameweeks

Live trigger status — evaluated on the fifteen on the pitch

The weekly cycle

Revision log — every time intelligence changed the plan

The sources, and how each one lies

0–0
Your record against me this season
Nothing has been played in 2026-27 yet, so this season is 0–0 and stays that way until Friday. Last season you were 4–0 against me, and every one of those four was about minutes or which club a player is at — never arithmetic. You read live sources; I read a file. That is the gap this page is trying to close.

What I need from you, gameweek to gameweek

This tool is only as good as what goes into it, and the parts that go stale fastest are the parts you can see and I cannot. Here is the exact input list, in priority order.

Every gameweek — five minutes, in this order

#WhatWhy it mattersWhere it goes
1Any team news you have that I do not — pressers, a manager quote, a player you saw benchedPredicted minutes beat every other input in the model. This is where you have a real edge over me.Conviction sliders, or tell me and I rebuild the XI data
2The actual gameweek scores — or just your rank and pointsLets me score the predictions and find which of my assumptions are failing, rather than which players got unlucky.The prediction ledger
3Your current squad, if it has drifted from what is in hereEvery risk number in this tool is computed against a specific fifteen.Copy link, or paste the fifteen names
4Your read on any club whose situation just changed — a sacking, a return, a formation switchRegime change is repriced by the market over six to ten weeks. That window is the edge and it opens the day the news lands.Belief switches and the club risk registers
5Chips used and transfers bankedOtherwise the plan drifts from your actual position.Chip planner

What I refresh at my end

  • Weekly: predicted XIs, injuries and suspensions, ownership, prices, the fixture ticker.
  • At GW6 and GW19: every table rebuilt on live 2026/27 data. The opponent-weakness numbers are currently 2025/26; three of last season's softest opponents were relegated, and the three promoted clubs carry a measured promotion prior rather than their own record.
  • From late February: blank and double gameweek tracking. Not December — I had that wrong. The first half is scheduled around the domestic cups and there was no blank or double in GW1–19 in either of the last two seasons; they arrive with the FA Cup fifth round and the run-in.
  • Never: the four-season base rates. The peak-chasing study, the absence-spell finding, the rotation result and the correlation measurement do not move.

The three moments that matter most

Thursday and Friday pressers, every week. Decide early, execute late. There is no price argument for moving early in the first half of the season, and waiting buys you role confirmation — which is worth far more, since early form explains only 14% of the rest of the season while role is observable in two gameweeks and stable.
GW6, after the international break. This is where differentials are judged — on role, not returns. If three of five have missed, the thesis was wrong; revert to continuity clubs and never add more differentials to recover.
GW15. Set a reminder now. All four first-set chips must be played by GW17 at the latest. They expire at the GW19 deadline — 13:30 GMT, Saturday 2 January — and an unplayed chip is worth exactly nothing. Do not hold any of them for a double gameweek: there was not one before GW19 in either of the last two seasons.

What I will tell you without being asked

  • When a belief you set is failing — with the evidence, not a vibe.
  • When one of my own predictions has been falsified. There are eleven on the record and I will report the tally every ten gameweeks whether it flatters me or not.
  • When the honest answer is to do nothing. Every worst case in FPL is made worse by activity.

Projections are computed from four seasons of gameweek data; scenario multipliers and start estimates are judgement, drawn from this session's research. Sources: FPL API · vaastav gameweek archives · Fantasy Football Scout · Opta/The Analyst · club officials and press conferences.