FPL Captain Picks — Who to Captain This Gameweek

Model captain rankings — 2026/27 launch preview

#	Player	Club	Pos	Price	Owned	Mins	Attack	Defence	xPts	Big haul	Captain pts
1	Haaland	MCI	FWD	£15.5m	73.5%	79	2.61	0.00	5.26	16%	10.52
2	Gabriel	ARS	DEF	£8.0m	26.9%	81	0.68	2.00	5.06	12%	10.12
3	B.Fernandes	MUN	MID	£12.0m	48.1%	80	1.91	0.62	5.04	16%	10.08
4	Palmer	CHE	MID	£9.5m	11.2%	80	2.09	0.39	4.60	14%	9.20
5	Mbeumo	MUN	MID	£8.0m	25.3%	64	1.95	0.25	4.21	12%	8.42
6	Isak	LIV	FWD	£9.0m	13.5%	65	1.63	0.00	4.05	10%	8.10
7	Wirtz	LIV	MID	£7.5m	14.4%	74	1.35	0.35	3.92	10%	7.84
8	Thiago	BRE	FWD	£8.0m	16.1%	80	1.75	0.08	3.90	11%	7.80
9	Gibbs-White	NFO	MID	£8.0m	11.6%	75	1.53	0.27	3.88	10%	7.76
10	Watkins	AVL	FWD	£8.0m	12.7%	70	1.74	0.00	3.85	10%	7.70

Same players, ranked by chance of a big haul

The list above is ordered by expected points. This one is ordered by the chance of a 10+ score. Where the two disagree it is nearly always the same story: a defender with a strong projection is a fine pick and a poor captain, because his good week is a clean sheet and a clean sheet does not win you a gameweek. We publish both and let you choose — over the 2025/26 walk-forward the haul ranking edged the points ranking, but across 31 gameweeks that is nowhere near enough to call it an edge, and we are not going to pretend otherwise.

#	Player	Club	Pos	Big haul	Bad / good week	xPts
1	B.Fernandes	MUN	MID	16%	2 – 12	5.04
2	Haaland	MCI	FWD	16%	2 – 13	5.26
3	Palmer	CHE	MID	14%	2 – 12	4.60
4	Mbeumo	MUN	MID	12%	1 – 11	4.21
5	Gabriel	ARS	DEF	12%	1 – 10	5.06
6	Thiago	BRE	FWD	11%	1 – 11	3.90
7	Watkins	AVL	FWD	10%	1 – 11	3.85
8	Gibbs-White	NFO	MID	10%	2 – 10	3.88
9	Wirtz	LIV	MID	10%	2 – 9	3.92
10	Isak	LIV	FWD	10%	2 – 9	4.05

Bad / good week = the score he lands on when it goes badly and when it goes well. One week in five is at or under the first number; one in ten beats the second.

Captaincy in rank terms — protect or chase

The table above ranks by expected points. But the armband is a rank decision — you gain on the field only for the points it isn't already on. Effective Ownership (EO) is the field's exposure (owned % + captaincy %); Captain Edge = xPts × (2 − EO/100) is the rank-points the armband gains over the field. Two archetypes fall out — using overall ownership as the field proxy until GW1 locks, then elite captaincy takes over automatically.

Protect rank · safeEO 73.5%

HaalandMCI

The field is most on him, so if he hauls you rise with the pack and a blank barely dents your rank. Captain pts 10.5 · big haul 16%.

Chase rank · aggressiveedge 8.7

PalmerCHE

Few rivals have him, so a haul is rank climbed past everyone who left him out — the highest Captain Edge on the shortlist, and the higher-variance bet. EO 11.2% · big haul 14%.

Player	Club	xPts	EO*	Big haul	Captain edge
Gabriel	ARS	5.06	26.9%	12%	8.76
Palmer	CHE	4.60	11.2%	14%	8.68
B.Fernandes	MUN	5.04	48.1%	16%	7.66
Isak	LIV	4.05	13.5%	10%	7.55
Mbeumo	MUN	4.21	25.3%	12%	7.35
Gibbs-White	NFO	3.88	11.6%	10%	7.31
Wirtz	LIV	3.92	14.4%	10%	7.28
Watkins	AVL	3.85	12.7%	10%	7.21
Thiago	BRE	3.90	16.1%	11%	7.17
Haaland	MCI	5.26	73.5%	16%	6.65

Captain Edge sorts by rank gained, not raw points — the safe pick sits low here by design, protecting rank rather than winning it. A defender can top this table on steady points and low ownership, but the armband is a ceiling bet, so the two picks above are drawn from midfielders and forwards — the positions that actually win a gameweek. *EO is overall ownership until GW1 locks; elite captaincy — what the top managers actually put the armband on — then takes over automatically.

Season-long proof: the model's armband vs the crowd's

Captaining the model's No. 1 every week of 2025/26 scored 187 points vs 176 for always captaining the most-owned player — ahead or level in 23 of 31 gameweeks, and marginally ahead on total points. Worth stating plainly: the model and the crowd landed on the same captain in 15 of those weeks — the default premium is usually the right call, and the model agrees. The crowd's pick beat the model outright 8 times, and every one of those weeks is in the log below — every call was made using only information available before that gameweek's deadline. Over 31 weeks a total that close is a coin flip, not an edge, and one gameweek can flip which side of it we land on. That is exactly why the armband table above now also publishes each player's chance of a big haul: on a sample this size the honest thing to give you is a better number to decide with, not a claim about who won.

What lands here every gameweek

One evergreen URL, headed "FPL Captain Picks — Gameweek {N}", rebuilt each gameweek. Four calls, each with a one-line model rationale — never a wall of names:

What 2025/26 proved: the ceiling almost never wears the crowd's armband

This is why the weekly surface publishes a safe pick and a differential rather than one name. Across the completed 2025/26 season, computed from the Official FPL API, the gameweek's highest scorer was a moving target the crowd's reflex pick rarely caught.

Safe vs differential — what the trade actually looked like

The template premiums bought a reliable floor: the two most-owned captain assets returned a double-figure haul in roughly a third of their matches. The best differentials cleared it far less often — but at a fraction of the ownership, so every hit moved rank against the field. That trade is exactly what the weekly expected-points × ownership table quantifies.

Bruno G. hit 10+ in 8 of 29 gameweeks at just 6.5% ownership — a differential floor rivalling the premiums' hit-rate. All figures final 2025/26.

Armband strategy — the quick rules