FPL 2026/27 · Opening Squad Guide

The model picks for one gameweek.
This guide looks further.

The model optimises for the gameweek in front of it: it takes that week's projections, applies the squad rules, and returns the best XI it can afford. It does not currently plan across a fixture run — a multi-gameweek planner is on the roadmap, and it will ship only if it beats the current one on an out-of-sample test, like every other change we have made.

That leaves a gap this guide fills. An opening squad is not really a one-week decision: you keep most of it for a month. So below, the model's projections sit alongside our own read of the first six fixtures — clearly labelled, so you always know which is which. From the model: per-player projections, the fixture table, the locked squad once published. From us: which projections to act on, chip timing, and what to watch before the deadline.

Fixture run: GW1–6

The shading and ordering below come from the model's own Dixon-Coles parameters — each cell shows the expected goals the model gives that team in that fixture, from the same file the player projections use. Teams are ranked by their total across the window. Home and away is shown as H / A.

Each cell shows the opponent, home or away, and the goals the model expects that team to score in that fixture. Shading is relative to the best and worst runs in this window; the final column totals expected goals across GW1-6. These come from the same Dixon-Coles parameters as the player projections, so the ordering is the model's view of the fixture run, not a hand-assigned difficulty rating. It reflects attacking output only - a good run for a striker is not automatically a good run for a defender.

Man City are the standout early-season pick. Bournemouth, Crystal Palace, and Coventry in the first three gameweeks is the kind of fixture run that makes premiums justified. GW4 at Manchester United is the first real test, and GW6 at Liverpool is the only genuinely difficult fixture in the block. City top the model's expected-goals table over these six fixtures. Our read: Haaland plus two City assets is the default opening structure.

Arsenal play Coventry at home in GW1, which flatters their opening record, then face a tricky GW2 at Villa and GW3 home to Chelsea. The run improves significantly in GW4–6. Arsenal defenders and Gyökeres are high-confidence picks over the six-week block, even if GW2–3 are not as easy as the raw fixtures suggest.

Liverpool have a manageable opening until GW6 when City visit Anfield. Newcastle's horror start — Liverpool at home then Spurs away — makes their GW3–6 run of Bournemouth, Leeds, Hull, Coventry look almost too easy in context. Their assets are worth holding past the rough opening.

The model's projections: premiums

The projection numbers are the model's output for GW1 — one gameweek only. Our commentary underneath each player is our read of that projection in context of the fixture run above.

The model's projections: mid-range

This is where the opening squad is usually won or lost. Getting two or three of these right is more impactful than the premium selections, because the differential here is larger and the budget freed up by a sub-premium pick is meaningful.

The model's projections: budget

Budget picks free up the funds for premiums. We target budget players who either have genuine clean sheet probability or have attacking returns from good fixtures at low cost. Avoid budget enablers who will ride the bench — always ask whether the saving is worth the trade-off in projected points.

Our captain plan: GW1–4

The optimiser produces one squad for one gameweek, not a rotation. The plan below is ours, built from the model's projections and the fixture table above. We do not armband-chase — the same player who was captain last week is not automatically the correct choice this week.

Our chip plan

Chip timing in FPL 2026/27 is most dependent on the blank and double gameweek calendar, which will not be announced until mid-season. These are our current expectations for when to hold or play each chip.

What we're watching

These are the situations we're monitoring before finalising the squad. We'll update this as pre-season news comes in.

• Gyökeres (ARS) Cold-start projection significantly underrates him given his 40+ goal season at Sporting. Needs Arsenal GW1 starting confirmation — if he's in the XI, he's in the squad.
• Donnarumma (MCI) If confirmed as City's #1 GK, his GW1–3 fixture run makes him arguably the best GK in the game for the opening block. Watch pre-season for lineup signals.
• Semenyo (MCI) Moved from Bournemouth to City. His role in City's system — starter or rotation — determines whether the £8.5m price tag is worth paying. Pre-season minutes are key.
• Mount (MUN) Budget option that makes the squad work financially. His injury record means this is a "monitor until GW0" situation — the price is right, the risk is the same it's always been.
• Calvert-Lewin (LEE) Projected 5.4 pts in GW1 at Nottingham Forest. Leeds' budget FWD slot may be worth it if he's fit and leads the line. Historically unavailable when you need him, but the price suggests they priced in the risk.
• Newcastle assets Hall and any Newcastle midfielder are worth monitoring once GW1 vs Liverpool is out of the way. Their GW3–6 run (Bournemouth, Leeds, Hull, Coventry) is one of the cleanest in the league.
• Transfer window (closes late Aug) Final week of the transfer window will confirm late arrivals and outgoings that change the value of many picks above. The model will re-project immediately after the window closes.

The model's squad for each gameweek is published before the deadline and graded afterwards. This guide is our commentary on top of it — where our read differs from the model, the record will show which was right.