How Fantasy EFL works

Fantasy EFL is the English Football League's official fantasy game, played across the Championship, League One and League Two. It asks you for three things every round: seven players, two clubs and one captain. This page is a short guide to those decisions and how the tools on this site help with them. It is not the official rulebook — for deadlines, eligibility and any rule changes, always check the official game.

The seven-player selection

You pick a small team, and "small" is the whole character of the game. Seven starters means every single pick is roughly a seventh of your score, so there is nowhere to hide a passenger. In a fifteen-man Premier League squad a bad fourth defender costs you a couple of points a week. Here, it costs you a seventh of the round.

The seven are one goalkeeper and six outfielders, in one of three formations:

• 1-2-2-2 — two defenders, two midfielders, two forwards
• 1-2-3-1 — two defenders, three midfielders, one forward
• 1-3-2-1 — three defenders, two midfielders, one forward

You may also take no more than two players from any one club. That limit is the part people underestimate. The top of any form-driven list clusters into the two or three clubs having a good month, so the seven best players are usually an illegal team, and the real question is which of them to give up. That is what Build your seven on the dashboard solves — it searches the legal combinations rather than taking the best player at each position and hoping.

Two more consequences, and both run against the instincts a Premier League fantasy game teaches you:

• Minutes matter more than form. Not a little more — measurably more. Across 83,698 real appearances, a player's average minutes over his last five rounds predicts his next round's points better than any points-based measure of form does. A player averaging under 30 minutes in his last five returns 0.59 points the following week; one averaging 75 or more returns 4.25. Check minutes before you check goals.
• Spreading risk is harder. With seven players you cannot cover every outcome, so a coherent squad — players from clubs with genuinely good fixtures — usually beats a scattering of individually clever picks.

The two club picks

Alongside your players you choose two clubs, and this is where most managers leave points behind. Club picks are quietly a fixture bet: you are backing a team's whole performance rather than an individual's, which means the things that decide them are more predictable than the things that decide a player's return. Form, defensive record and the difficulty of the next match do most of the work.

They are also the easiest part of the game to set once in August and never look at again. Don't. A club with three straight home games against the bottom of its division is a different proposition from the same club a month later. The club picker ranks all 72 on form, goals scored and conceded, clean sheets, home and away records and the fixtures ahead — and shows the reasoning, so you can override it when you know something it does not.

The captain

Your captain's return counts for more than anyone else's, which makes it the single highest-variance decision you make each round. The instinct is to captain the best player you own. The better habit is to captain the best situation you own: a strong player, at home, against a side conceding regularly, who is certain to start.

That last clause does more work than any other. A captain who is rested after a midweek game does not score you half of nothing — he scores you nothing, twice. Check the fixture ticker for short turnarounds before you commit the armband.

Do not assume the armband belongs to a forward. That is a Premier League habit and this game does not reward it. Measured across those same 83,698 appearances, mean points per appearance run defenders 4.19, goalkeepers 4.08, midfielders 3.88, forwards 3.14. The forward is the worst-scoring position in Fantasy EFL, because defenders are paid for clean sheets, clearances, blocks and tackles, and a goalkeeper's goal is worth ten points. A defender at home to the bottom of the division is very often the right captain.

How points are scored

This page used to decline to print the scoring table, on the grounds that guessing at another game's rules in public is worse than saying "look it up". It is no longer a guess. The table below was checked by recomputing the score of 83,698 real player-round records — 35 rounds of a completed season, published by the official game — from their raw statistics and comparing with the official points figure. 83,688 of them reproduce exactly: 99.99%. The ten misses are all off by one, which is what post-match statistical corrections look like.

Tariffs can change between seasons, and the official game is the authority. But this is accurate enough to plan with, and accurate enough that the models on this site score players in the game's real currency rather than an approximation of it.

Three things in that table do most of the work, and none of them is a goal.

• Interceptions pay two points each, and only midfielders get them. It is the most valuable repeatable statistic in the game. A midfielder making four interceptions has out-scored a forward's shot on target eight times over, in a way that never shows up in a highlights package.
• Defenders get paid for defending — clearances, blocks and tackles on top of the clean sheet. That, plus a goal being worth seven, is why they out-score everyone.
• A goalkeeper's goal is worth ten points, which will almost never happen and is worth knowing anyway.

The player finder shows what each statistic was worth next to the statistic itself, and switches its columns to match the position you filter to — a defender's clearances, a midfielder's interceptions.

Fixture difficulty

Every fixture on this site carries a rating from 1 (most favourable) to 5 (toughest). It is built from the opponent's points per game, their defensive record and their attacking record, adjusted for home advantage — and it is calculated within a division. A rating of 2 in League Two means "comfortably easier than average for League Two", not "easier than a Championship fixture".

That division-local approach is deliberate, and it is what makes cross-division comparison possible at all. If ratings were calculated across all 72 clubs at once, every League Two fixture would look easy, every Championship fixture would look hard, and the ticker would just be telling you which division you were reading.

Fixture ratings are a modelled guide, not a guarantee. They are built from results so far and cannot see a manager change, an injury crisis, a cup replay or a club that has quietly been much better than its table position for a month.

Short-term form, starts and minutes

Form on this site means the last five rounds, and nothing longer. Over a 46-game season a player's total points tell you who has been good since August; the last five tell you who is good now, and only one of those is a reason to pick someone this weekend.

Form is read alongside minutes rather than instead of them. Twelve points in two appearances is a different signal from twelve in five, and a player averaging 60 minutes is a different proposition from one who plays 90 every week. The finder shows points per appearance, starts, minutes and availability side by side so the three can be read together.

Availability

The cheapest points in any fantasy game are the ones you don't lose to a player who was never going to play. Every player on this site carries a status — available, doubtful, injured, suspended or unavailable — and the modelled rating treats it as a multiplier rather than a minor deduction. An injured player with perfect form is not a slightly worse pick; he is not a pick.

Bookings and suspension

The EFL suspends a player for accumulating yellow cards, on a ladder with deadlines. The count is cumulative and does not reset when a ban is served — a player who sat out at five is next in trouble at ten, not at five again.

A player one booking away is flagged on the player finder and on the pick cards, and the finder can filter to or away from them.

It does not change anybody's rating, on purpose. A ban costs the round after the booking, not the round you are picking — so marking a player down today for a match he will miss next week would be scoring the wrong week. The availability multiplier already handles a player who is suspended now. This is a planning signal instead: he may be missing next round, and how much that costs you depends on how freely the official game lets you change your side, which is a rule this site does not claim to know. Check it before you build a squad around someone on four bookings.

Nor is it a prediction. It reports how many bookings a player is from a ban. It does not estimate his chance of picking one up.

Comparing across three divisions

This is the skill Fantasy EFL asks for that no single-division game does. A forward scoring freely in League Two and a Championship midfielder with three assists in four are genuinely hard to compare, and the honest answer is that raw totals cannot do it.

So every input in the models here is normalised within a player's own division before anything is ranked: form against players in the same division and position, fixture difficulty against clubs in the same division, club form against the same twenty-four teams. What is compared across divisions is a player's standing, not their raw numbers.

Differentials

A differential normally means a pick few other managers own, and Fantasy EFL splits neatly in two on this.

For clubs, ownership is real and published. The official game states what percentage of managers has picked each club, and where the data source in use supplies it, the club picker shows it as a column and lets you sort by it. A strong club that few managers have taken is the closest thing to a genuine differential this game offers, and it is the easier of the two club picks to get an edge on precisely because most managers set theirs in August and forget.

For players, nothing public publishes it — so this site shows no player ownership column and does not invent one. What the finder offers instead is a form differential: an editorial, modelled score combining strong recent output with how much attention a player's club tends to get — its division, and how near the top of it the club sits. It is a prompt to go and look at someone, not evidence that they are lightly owned, and it is labelled as modelled wherever it appears.

How the recommendations are built

Nothing on this site is a black box. Two weighted models do all the work, and both are in one file with the weights at the top of it.

Player rating

These weights are not picked by feel. They come from a walk-forward test on the same 83,698-appearance dataset: for each of 52,158 player-rounds, build a predictor from earlier rounds only and see how well it correlates with the points actually scored in the next round.

That last row is worth a sentence. Points per 90 is a natural-looking measure and it is very nearly useless for picking, because it flatters a substitute who scores in a twenty-minute cameo. Form on this site is always points per appearance.

• Starts and minutes (34%) — the strongest single signal there is
• Recent form (26%) — points per appearance in the last five rounds
• Output (16%) — everything the player has been paid for, valued at the real tariff above and expressed per 90 minutes
• Next fixture (16%) — the 1-5 rating above
• Home advantage (8%)
• Availability — applied as a multiplier over the total, not as another slice

The last two are then scaled by position, because a flat number is wrong for both. Home advantage is worth +0.7% to a goalkeeper and +14.3% to a forward. Fixture difficulty runs the other way: across all positions, points move about 17% between the easiest and hardest fixture bands, but a goalkeeper's or defender's clean-sheet rate moves from 36.3% to 19.0% — nearly half, and a clean sheet is five points. The fixture is the pick for a defender; for a forward it is one input among several.

Club rating

• Recent form (28%) — league points in the last five
• Upcoming fixtures (28%) — mean difficulty of the next three
• Goals conceded (20%) — in the last five
• Goals scored (17%) — in the last five
• Home matches (7%) — how many of the next three are at home

Both produce a score out of 100 and a plain-English sentence assembled from whichever factors contributed most — "strong recent form, a favourable home fixture and three consecutive starts" is a summary of the arithmetic, not a caption written over it. Where a score is shown, the reasoning is shown with it.

These are ratings, not predictions. They rank the options in front of you against stated criteria using results that have already happened. They cannot see team news, weather, a dressing-room row or a manager who has decided to rest half a side before a cup tie — which is exactly the sort of thing you know and they do not. Use them as a shortlist and then apply what you know.

You do not have to take that on trust. Every round, before the deadline, the model's seven players, captain and two clubs are written down, and after the round they are graded against what a random seven scored, what "just pick whoever has been scoring" scored, and what the best possible seven scored. The whole record — including the bad weeks — is on the model's record.

Where the numbers come from

Loading data source…

Whatever the source, the fixture ratings, player ratings and club ratings are calculated here, by the models described above. They are never supplied by anyone else and are never official.

Is the feed actually working?

The paragraph above reports what this page's own data load returned a moment ago. For a deeper answer — every document the app depends on, fetched fresh, with the fields it needs checked one by one — run the check below. It is a button rather than something that happens automatically because it deliberately skips every cache and asks the official game's servers for all three documents; firing that on every page view would make this site the impolite client the caching exists to prevent.

It reads three public documents and sends no account, cookie or key of any kind. The same report is available as JSON at /api/efl/health.

Check the feed now

Start here

• This round's picks — seven cards, one per decision, and a legal seven built for you
• Fixture ticker — six rounds ahead for all 72 clubs
• Player finder — filter and rank every player
• Club picker — compare the two club selections