XI Predictor guide

Gameweek 38

Picks, captains, free hit XI, value plays, differentials and per-position rankings.

Live tools — always up to date

The headline picks

Top captain
Haaland
6.7
xPts GW38
Man City · FWD · £14.7m · 64.4% owned
Fixture:
Aston Villa (H)
Best value
Darlow
3.5
xPts GW38
Leeds · GKP · £4.0m · 6.8% owned
Fixture:
West Ham (A)
Top differential
Bruno G.
4.6
xPts GW38
Newcastle · MID · £6.9m · 6.2% owned
Fixture:
Fulham (A)

Browse the Gameweek 38 guides

Frequently asked questions

Who is the best captain for Gameweek 38?

Our model's top captain pick is Haaland (Man City) with 6.7 expected points. Strong alternatives include B.Fernandes and Gabriel.

Should I play my Free Hit in Gameweek 38?

Free Hit is best when your normal squad has many blanks or weak fixtures. Our optimised XI for the gameweek projects 54.2 expected points before captaincy.

What's a good differential for Gameweek 38?

Bruno G. (Newcastle, 6.2% owned, 4.6 xPts) leads our differential picks under 10% ownership.

How does XI Predictor calculate these picks?

Each player is given an Expected Points (xPts) score blending recent form, expected goal involvements (xG + xA), fixture difficulty, and expected minutes. Picks are deterministic from the data snapshot at publish time.

Glossary
xPts
Expected points: our model's forecast for the player's next gameweek score.
FDR
Fixture difficulty rating, 1 (easiest) to 5 (hardest).
xG / xA / xGI
Expected goals, expected assists, and the sum (expected goal involvements).
ICT
The Influence-Creativity-Threat composite published with official data.
Ownership
Percent of managers who currently own the player.
Form
Recent points-per-match average over the last several matches.
xPts/£m
Expected points produced per £1m of price — our value score.
Snapshot taken 20 May 2026
Picks are deterministic from a snapshot of player news, prices, fixtures, and form at publish time. The model blends Expected Points (xPts) with fixture difficulty, expected minutes, and bonus likelihood. Read the full methodology.