How to use this

Every week a model looks at every Premier League player — recent form, how good or bad their opponent is, whether they take the penalties and corners, how likely they are to play the full 90 — and works out how many FPL points each one should score. It then builds the strongest possible team from those numbers.

You don't have to take its word for anything. Every prediction is locked in before the deadline and later scored against what really happened, in public. You can pull up that track record any time — see the Model Scoreboard below.

Most weeks you're not building a team from scratch — you're making one or two transfers. That's exactly what this is for.

The pitch shows the eleven the model rates most highly this week. Tap any player and a card slides up explaining, in plain English, why he's picked — what he brings, why the fixture suits him, and what to expect (a goal, a clean sheet, a big haul). If a highly-rated player is someone you don't own, that's your shortlist for a transfer. The reasoning tells you whether he's a nailed-on starter in great form or more of a punt.

Your captain scores double, so it's the biggest call of the week. Under the pitch are three captain options — tap any of them to see the reasoning behind it:

The Top Pick — the safest bet, the player most likely to deliver. If you just want the sensible answer, take this and move on.

The Differential — a strong player few managers own. If it comes off you leap past everyone who picked the obvious name. Use it when you're chasing.

The Defensive Pick — a defender or keeper with a strong clean-sheet chance. Lower ceiling, higher floor, for cautious weeks.

The eleven is shown in formation with the bench below. The copper C marks the recommended captain. Tap anyone for the full breakdown.

The one number worth knowing is xP — "expected points". It's the player's average score if this weekend were replayed a thousand times. An xP of 6 doesn't promise 6 points; it means the odds lean strongly toward a good return.

At the very start of a new season there's no current form to go on yet — new signings, new managers, promoted teams. For the first couple of gameweeks the model leans on last season's numbers, so treat it as a strong shortlist rather than gospel, and lean on your own read of who's actually starting.

By around Gameweek 5 it has enough of the new season to trust properly. The dashboard shows a small note during this early window so you always know where you stand.

The row of chips under the header lets you rewind. Tap any past week to see the exact team the model named before that deadline, what each player actually scored, and whether the eleven beat the average FPL manager.

The Model Scoreboard sums it up in one line: the model's points per week versus the average manager's, and how many weeks it came out on top. Some weeks it wins, some it doesn't — but nothing is ever hidden or rewritten, so you can judge for yourself how much to trust it. Each new season the record starts fresh from zero.

1. Open the dashboard a day or two before the deadline.
2. Tap the highly-rated players you don't own — that's your transfer shortlist.
3. Compare the three captain picks against your gut, and choose one.
4. Make your move before the deadline.
5. After the gameweek, open the Time Machine to see how it went.

xP — expected points, the model's best estimate.
Haul — a big single-week score (9+ points).
Differential — a strong player few managers own.
Clean sheet — no goals conceded; big points for defenders and keepers.
Owned by — the share of managers who have the player.
Form — average points over the last few matches.