Fantasy Premier League
Differentials worth owning
A player nobody owns is not a differential on his own — most of them are players nobody owns for a reason. Every row here clears an ownership threshold and a minutes floor, is ranked on the model's own currency, and says how much of its edge is the projection and how much is only the field's absence.
James Trafford
GKP · LEE · 5.0m · owned by 5.5%
+1.26
Edge
That is 9.1 projected points over the next 3 rounds, which is what the board under it is ranked on.
Gameweek 6 · 3-round window · feed of 01/10 11:39
James Trafford is the best-projected player in the game owned by under 10%: 9.1 over 3 rounds at 5.0m, held by 5.5%.
He is behind David Raya Martín, the most-owned player in his position at 42%, on projection by −0.52, and +1.78 of his edge is the field he is not in. That is what a player nobody owns for a reason looks like from here: the ownership is the whole of the case.
Own
5.5%
The share of managers holding him, against the 10% gate this board opens at.
Proj edge
−0.52
What the model rates him over the most-owned player in his position. Positive is football; negative means the only case for him is that nobody owns him.
Own edge
+1.78
What the field's absence is worth at what the template is worth. The two halves add up to his whole edge.
2 low-owned players are left off because the minutes model puts them under 45% to start.
See the minutes board →The top five on this board
The board as your controls leave it — the same five that open the table below.
-
Ranked 1 James Trafford9.1 Ahead
- Own
- 5.5%
- Proj edge
- −0.52
- Own edge
- +1.78
Against David Raya Martín
3.7- ARS (A)
- MUN (H)
- SUN (A)
-
Ranked 2 Senne Lammens8.9 Ahead
- Own
- 9.9%
- Proj edge
- −0.69
- Own edge
- +1.56
Against David Raya Martín
2.3- TOT (H)
- LEE (A)
- BOU (H)
-
Ranked 3 Alisson Becker8.7 Ahead
- Own
- 4.1%
- Proj edge
- −0.93
- Own edge
- +1.85
Against David Raya Martín
3.0- MCI (H)
- BRE (A)
- BHA (H)
-
Ranked 4 Bernd Leno8.4 Ahead
- Own
- 3.4%
- Proj edge
- −1.23
- Own edge
- +1.88
Against David Raya Martín
3.0- IPS (A)
- HUL (H)
- COV (A)
-
Ranked 5 Gianluigi Donnarumma8.4 Ahead
- Own
- 7.6%
- Proj edge
- −1.24
- Own edge
- +1.68
Against David Raya Martín
2.7- LIV (A)
- IPS (H)
- AVL (A)
- Proj edge
- −0.52
- Own edge
- +1.78
- Pts
- 3.7
Against David Raya Martín
- ARS (A)
- MUN (H)
- SUN (A)
- Proj edge
- −0.69
- Own edge
- +1.56
- Pts
- 3.8
Against David Raya Martín
- TOT (H)
- LEE (A)
- BOU (H)
- Proj edge
- −0.93
- Own edge
- +1.85
- Pts
- 3.6
Against David Raya Martín
- MCI (H)
- BRE (A)
- BHA (H)
- Proj edge
- −1.23
- Own edge
- +1.88
- Pts
- 3.3
Against David Raya Martín
- IPS (A)
- HUL (H)
- COV (A)
- Proj edge
- −1.24
- Own edge
- +1.68
- Pts
- 3.3
Against David Raya Martín
- LIV (A)
- IPS (H)
- AVL (A)
- Proj edge
- −1.43
- Own edge
- +1.75
- Pts
- 3.3
Against David Raya Martín
- AVL (A)
- LIV (H)
- HUL (A)
- Proj edge
- −1.50
- Own edge
- +1.60
- Pts
- 3.3
Against David Raya Martín
- HUL (A)
- CHE (H)
- ARS (A)
- Proj edge
- −1.75
- Own edge
- +2.00
- Pts
- 3.4
Against David Raya Martín
- NEW (H)
- TOT (A)
- FUL (H)
- Proj edge
- −1.91
- Own edge
- +1.97
- Pts
- 3.1
Against David Raya Martín
- COV (A)
- AVL (H)
- CRY (A)
- Proj edge
- −1.96
- Own edge
- +1.90
- Pts
- 3.1
Against David Raya Martín
- CHE (A)
- SUN (H)
- MUN (A)
- Proj edge
- −2.14
- Own edge
- +1.82
- Pts
- 3.1
Against David Raya Martín
- BOU (H)
- EVE (A)
- TOT (H)
- Proj edge
- −2.43
- Own edge
- +1.93
- Pts
- 2.9
Against David Raya Martín
- BHA (H)
- BOU (A)
- LEE (H)
- Proj edge
- −2.45
- Own edge
- +1.95
- Pts
- 3.0
Against David Raya Martín
- BRE (H)
- NEW (A)
- MCI (H)
- Proj edge
- −2.94
- Own edge
- +2.01
- Pts
- 2.8
Against David Raya Martín
- FUL (H)
- MCI (A)
- NFO (H)
What an edge is here, and what it is measured against
Every row is owned by under 10% of the game. The threshold is the page's premise rather than a sort: the ranking under it is the weighted score for the rounds ahead, which is a unit this model is measured in.
The two edge columns are one figure split in two, under the rule this app already trades points against ownership by — points times one minus 0.5 times ownership, which is the captain shortlist's own aggressive setting. They add up to the whole edge exactly, so a reader can check the pair rather than trust it.
The template is the most-owned player the game has in that position, taken off the whole board rather than off the filtered rows — a differential is measured against the field, not against the least contrarian row left on the screen. A position the game has published nobody's ownership for has no template, and both columns print a dash rather than a nought.
The minutes floor is the projection's own start probability at 45%. It is the largest way a list like this goes wrong, so it is a gate rather than a column — and nothing on this site backtests the ownership half of the split.
A player the game has published no ownership figure for is left off rather than counted as owned by nobody. Missing is not nought, and the one place that mistake does the most damage is a board about ownership.
Every ranking here is in expected points or in the weighted score for the rounds ahead, which are the units this model is measured in. The value column is those figures restated over a price — it is a way of scanning the board, and nothing on this site backtests it.
Bonus points are projected nowhere on this site, so they are in none of these figures.
The board is every player the game would let you field, the flagged included — whether a doubtful player is worth picking is a judgement the projection makes with his published chance of playing, not one made by hiding him.
What is measured here, and what is not
-
Measured The projected points
Replayed against gameweek 2 after it was played: an average miss of 1.617 points a player, against 1.842 for “predict what he has averaged so far”, over 456 players. Recorded Sunday, 6 September 2026 and refreshed by hand, so the date is part of the claim.