Predictions · Gameweek 1

What we say will happen

These are FPL projections from our own fixture model, filed before kick-off and scored once the match is over.

Two Poissons: each club's expected goals conceded per 90, stretched by FPL's own fixture difficulty, then every 0-6 scoreline added up.

Goals by the two sides are treated as independent. That flatters both-to-score in open games and understates it in sieges. Probabilities, not odds. FPL projections from our own model. The squad-level record lives on /accuracy.

This page follows Gameweek 1 from the live FPL fixture list — when this week is fully played it will roll to the next one, without an edit. · Phases as of 20 Aug, 08:57 (build time), from published kick-off times.

DeadlineFriday 21 August at 18:30

What we project

Result split, both teams to score, clean sheets, team goal counts, and the top attackers' chances of a goal or assist. Where it helps, we say none, exactly one, or two or more instead of hiding everything inside one percentage.

What we do not claim

A full coupon can list 200+ lines for one match. We do not copy handicaps, exact scores, half-time markets, first scorer, cards, or corners. We answer only the questions our FPL data can support, and we answer them as probabilities, not bets.

Upcoming

8 fixtures, earliest kick-off first

EVE v CRY

Sat 22 Aug, 15:00 UK

Match sheet

Result projection

EVE 36%Draw 25%CRY 39%

Match climate
Medium
2.9 goals
0-1 22% · 2-3 46% · 4+ 32%
Both to score
58%
Clean sheet
EVE 23%
CRY 24%
To score
EVE 76%
CRY 77%
EVE: 76% to score at least once. 34% exactly one. 41% two or more.
CRY: 77% to score at least once. 34% exactly one. 43% two or more.

EVE · Attacking return

  • Beto41%goal 40% · assist 1% · rotation riskAssists: 1% at least one. 1% exactly one. 0% two or more. Most likely: none.Goals: 40% at least one. 31% exactly one. 9% two or more. Most likely: none.
  • Barry34%goal 33% · assist 2% · probableAssists: 2% at least one. 2% exactly one. 0% two or more. Most likely: none.Goals: 33% at least one. 27% exactly one. 6% two or more. Most likely: none.

CRY · Attacking return

  • Mateta47%goal 44% · assist 5% · probableAssists: 5% at least one. 5% exactly one. 0% two or more. Most likely: none.Goals: 44% at least one. 33% exactly one. 12% two or more. Most likely: none.
  • Sarr39%goal 36% · assist 5% · probableAssists: 5% at least one. 5% exactly one. 0% two or more. Most likely: none.Goals: 36% at least one. 29% exactly one. 7% two or more. Most likely: none.

Result split sums to 100% — the remaining 0% sits in scorelines beyond 6-6, which the grid truncates. Home or draw is the two figures added.

These are probabilities from our model, not a confirmation and not a betting tip. Football is noisy. We will be wrong. We score ourselves in public. Player count splits use the same model as the headline percentages, but only fixture-level numbers are filed in the public record today.

IPS v SUN

Sat 22 Aug, 15:00 UK

Match sheet

Result projection

IPS 36%Draw 28%SUN 36%

Match climate
Medium
2.4 goals
0-1 31% · 2-3 47% · 4+ 22%
Both to score
48%
Clean sheet
IPS 31%
SUN 30%
To score
IPS 70%
SUN 69%
IPS: 70% to score at least once. 36% exactly one. 34% two or more.
SUN: 69% to score at least once. 36% exactly one. 33% two or more.

IPS · Attacking return

  • Lukić20%goal 11% · assist 10% · rotation riskAssists: 10% at least one. 9% exactly one. 0% two or more. Most likely: none.Goals: 11% at least one. 11% exactly one. 1% two or more. Most likely: none.
  • Diop10%goal 9% · assist 1% · unlikelyAssists: 1% at least one. 1% exactly one. 0% two or more. Most likely: none.Goals: 9% at least one. 9% exactly one. 0% two or more. Most likely: none.

SUN · Attacking return

  • Isidor35%goal 33% · assist 3% · rotation riskAssists: 3% at least one. 3% exactly one. 0% two or more. Most likely: none.Goals: 33% at least one. 27% exactly one. 6% two or more. Most likely: none.
  • E.Le Fée30%goal 15% · assist 17% · likely starterAssists: 17% at least one. 16% exactly one. 2% two or more. Most likely: none.Goals: 15% at least one. 14% exactly one. 1% two or more. Most likely: none.

Result split sums to 100% — the remaining 0% sits in scorelines beyond 6-6, which the grid truncates. Home or draw is the two figures added.

These are probabilities from our model, not a confirmation and not a betting tip. Football is noisy. We will be wrong. We score ourselves in public. Player count splits use the same model as the headline percentages, but only fixture-level numbers are filed in the public record today.

NFO v LEE

Sat 22 Aug, 15:00 UK

Match sheet

Result projection

NFO 42%Draw 25%LEE 32%

Match climate
Medium
2.8 goals
0-1 24% · 2-3 46% · 4+ 30%
Both to score
56%
Clean sheet
NFO 28%
LEE 23%
To score
NFO 77%
LEE 72%
NFO: 77% to score at least once. 34% exactly one. 44% two or more.
LEE: 72% to score at least once. 36% exactly one. 36% two or more.

NFO · Attacking return

  • Wood36%goal 34% · assist 3% · rotation riskAssists: 3% at least one. 3% exactly one. 0% two or more. Most likely: none.Goals: 34% at least one. 27% exactly one. 7% two or more. Most likely: none.
  • Gibbs-White34%goal 28% · assist 8% · likely starterAssists: 8% at least one. 8% exactly one. 0% two or more. Most likely: none.Goals: 28% at least one. 24% exactly one. 4% two or more. Most likely: none.

LEE · Attacking return

  • Nmecha52%goal 45% · assist 12% · rotation riskAssists: 12% at least one. 12% exactly one. 1% two or more. Most likely: none.Goals: 45% at least one. 33% exactly one. 12% two or more. Most likely: none.
  • Calvert-Lewin42%goal 40% · assist 3% · probableAssists: 3% at least one. 3% exactly one. 0% two or more. Most likely: none.Goals: 40% at least one. 31% exactly one. 9% two or more. Most likely: none.

Result split sums to 100% — the remaining 0% sits in scorelines beyond 6-6, which the grid truncates. Home or draw is the two figures added.

These are probabilities from our model, not a confirmation and not a betting tip. Football is noisy. We will be wrong. We score ourselves in public. Player count splits use the same model as the headline percentages, but only fixture-level numbers are filed in the public record today.

BRE v TOT

Sat 22 Aug, 17:30 UK

Match sheet

Result projection

BRE 37%Draw 25%TOT 38%

Match climate
Medium
3.0 goals
0-1 21% · 2-3 45% · 4+ 34%
Both to score
59%
Clean sheet
BRE 23%
TOT 23%
To score
BRE 77%
TOT 77%
BRE: 77% to score at least once. 34% exactly one. 43% two or more.
TOT: 77% to score at least once. 34% exactly one. 44% two or more.

BRE · Attacking return

  • Thiago45%goal 42% · assist 5% · likely starterAssists: 5% at least one. 5% exactly one. 0% two or more. Most likely: none.Goals: 42% at least one. 32% exactly one. 11% two or more. Most likely: none.
  • O.Dango37%goal 26% · assist 15% · probableAssists: 15% at least one. 13% exactly one. 1% two or more. Most likely: none.Goals: 26% at least one. 22% exactly one. 4% two or more. Most likely: none.

TOT · Attacking return

  • Richarlison37%goal 33% · assist 6% · rotation riskAssists: 6% at least one. 6% exactly one. 0% two or more. Most likely: none.Goals: 33% at least one. 27% exactly one. 6% two or more. Most likely: none.
  • Tel28%goal 15% · assist 15% · rotation riskAssists: 15% at least one. 14% exactly one. 1% two or more. Most likely: none.Goals: 15% at least one. 14% exactly one. 1% two or more. Most likely: none.

Result split sums to 100% — the remaining 0% sits in scorelines beyond 6-6, which the grid truncates. Home or draw is the two figures added.

These are probabilities from our model, not a confirmation and not a betting tip. Football is noisy. We will be wrong. We score ourselves in public. Player count splits use the same model as the headline percentages, but only fixture-level numbers are filed in the public record today.

BHA v AVL

Sun 23 Aug, 14:00 UK

Match sheet

Result projection

BHA 38%Draw 26%AVL 36%

Match climate
Medium
2.7 goals
0-1 25% · 2-3 47% · 4+ 29%
Both to score
55%
Clean sheet
BHA 27%
AVL 25%
To score
BHA 75%
AVL 73%
BHA: 75% to score at least once. 35% exactly one. 40% two or more.
AVL: 73% to score at least once. 35% exactly one. 38% two or more.

BHA · Attacking return

  • Hinshelwood31%goal 22% · assist 11% · rotation riskAssists: 11% at least one. 11% exactly one. 1% two or more. Most likely: none.Goals: 22% at least one. 20% exactly one. 3% two or more. Most likely: none.
  • De Cuyper29%goal 9% · assist 22% · rotation riskAssists: 22% at least one. 19% exactly one. 3% two or more. Most likely: none.Goals: 9% at least one. 9% exactly one. 0% two or more. Most likely: none.

AVL · Attacking return

  • Watkins40%goal 38% · assist 4% · likely starterAssists: 4% at least one. 4% exactly one. 0% two or more. Most likely: none.Goals: 38% at least one. 30% exactly one. 8% two or more. Most likely: none.
  • Garnacho40%goal 22% · assist 23% · rotation riskAssists: 23% at least one. 20% exactly one. 3% two or more. Most likely: none.Goals: 22% at least one. 19% exactly one. 3% two or more. Most likely: none.

Result split sums to 100% — the remaining 0% sits in scorelines beyond 6-6, which the grid truncates. Home or draw is the two figures added.

These are probabilities from our model, not a confirmation and not a betting tip. Football is noisy. We will be wrong. We score ourselves in public. Player count splits use the same model as the headline percentages, but only fixture-level numbers are filed in the public record today.

MCI v BOU

Sun 23 Aug, 14:00 UK

Match sheet

Result projection

MCI 54%Draw 22%BOU 23%

Match climate
High
3.1 goals
0-1 18% · 2-3 44% · 4+ 38%
Both to score
59%
Clean sheet
MCI 31%
BOU 15%
To score
MCI 85%
BOU 69%
MCI: 85% to score at least once. 28% exactly one. 57% two or more.
BOU: 69% to score at least once. 36% exactly one. 33% two or more.

MCI · Attacking return

  • Haaland56%goal 52% · assist 7% · likely starterAssists: 7% at least one. 7% exactly one. 0% two or more. Most likely: none.Goals: 52% at least one. 35% exactly one. 17% two or more. Most likely: none.
  • Cherki49%goal 20% · assist 37% · rotation riskAssists: 37% at least one. 29% exactly one. 8% two or more. Most likely: none.Goals: 20% at least one. 18% exactly one. 2% two or more. Most likely: none.

BOU · Attacking return

  • Brooks45%goal 31% · assist 20% · rotation riskAssists: 20% at least one. 18% exactly one. 2% two or more. Most likely: none.Goals: 31% at least one. 26% exactly one. 6% two or more. Most likely: none.
  • Tavernier32%goal 23% · assist 12% · probableAssists: 12% at least one. 11% exactly one. 1% two or more. Most likely: none.Goals: 23% at least one. 20% exactly one. 3% two or more. Most likely: none.

Result split sums to 100% — the remaining 0% sits in scorelines beyond 6-6, which the grid truncates. Home or draw is the two figures added.

These are probabilities from our model, not a confirmation and not a betting tip. Football is noisy. We will be wrong. We score ourselves in public. Player count splits use the same model as the headline percentages, but only fixture-level numbers are filed in the public record today.

NEW v LIV

Sun 23 Aug, 16:30 UK

Match sheet

Result projection

NEW 29%Draw 25%LIV 45%

Match climate
Medium
2.7 goals
0-1 24% · 2-3 46% · 4+ 29%
Both to score
55%
Clean sheet
NEW 21%
LIV 30%
To score
NEW 70%
LIV 79%
NEW: 70% to score at least once. 36% exactly one. 33% two or more.
LIV: 79% to score at least once. 33% exactly one. 46% two or more.

NEW · Attacking return

  • Osula40%goal 36% · assist 6% · rotation riskAssists: 6% at least one. 6% exactly one. 0% two or more. Most likely: none.Goals: 36% at least one. 29% exactly one. 7% two or more. Most likely: none.
  • Barnes34%goal 26% · assist 11% · rotation riskAssists: 11% at least one. 11% exactly one. 1% two or more. Most likely: none.Goals: 26% at least one. 22% exactly one. 4% two or more. Most likely: none.

LIV · Attacking return

  • Ngumoha42%goal 20% · assist 28% · rotation riskAssists: 28% at least one. 24% exactly one. 4% two or more. Most likely: none.Goals: 20% at least one. 18% exactly one. 2% two or more. Most likely: none.
  • Wirtz36%goal 22% · assist 17% · probableAssists: 17% at least one. 16% exactly one. 2% two or more. Most likely: none.Goals: 22% at least one. 20% exactly one. 3% two or more. Most likely: none.

Result split sums to 100% — the remaining 0% sits in scorelines beyond 6-6, which the grid truncates. Home or draw is the two figures added.

These are probabilities from our model, not a confirmation and not a betting tip. Football is noisy. We will be wrong. We score ourselves in public. Player count splits use the same model as the headline percentages, but only fixture-level numbers are filed in the public record today.

FUL v CHE

Mon 24 Aug, 20:00 UK

Match sheet

Result projection

FUL 33%Draw 24%CHE 42%

Match climate
Medium
3.0 goals
0-1 21% · 2-3 45% · 4+ 34%
Both to score
59%
Clean sheet
FUL 21%
CHE 25%
To score
FUL 75%
CHE 80%
FUL: 75% to score at least once. 35% exactly one. 40% two or more.
CHE: 80% to score at least once. 32% exactly one. 47% two or more.

FUL · Attacking return

  • Bobb25%goal 11% · assist 16% · rotation riskAssists: 16% at least one. 14% exactly one. 1% two or more. Most likely: none.Goals: 11% at least one. 11% exactly one. 1% two or more. Most likely: none.
  • Muniz25%goal 24% · assist 1% · rotation riskAssists: 1% at least one. 1% exactly one. 0% two or more. Most likely: none.Goals: 24% at least one. 21% exactly one. 3% two or more. Most likely: none.

CHE · Attacking return

  • Palmer42%goal 36% · assist 10% · probableAssists: 10% at least one. 9% exactly one. 1% two or more. Most likely: none.Goals: 36% at least one. 28% exactly one. 7% two or more. Most likely: none.
  • João Pedro42%goal 38% · assist 6% · probableAssists: 6% at least one. 6% exactly one. 0% two or more. Most likely: none.Goals: 38% at least one. 30% exactly one. 8% two or more. Most likely: none.

Result split sums to 100% — the remaining 0% sits in scorelines beyond 6-6, which the grid truncates. Home or draw is the two figures added.

These are probabilities from our model, not a confirmation and not a betting tip. Football is noisy. We will be wrong. We score ourselves in public. Player count splits use the same model as the headline percentages, but only fixture-level numbers are filed in the public record today.

Not projected (2)

What we said, and what happened

Squad record → /accuracy

Every probability below is the one filed before kick-off, read back from the log. None of it is recomputed now that the result is known.

No match projections have been scored yet. This section stays empty until a gameweek has been played.

Read the verdicts carefully. One match cannot prove a probability right or wrong. Only the pooled Brier score and the calibration table below can do that. With us means the side of the call we leaned toward happened; against us means it did not; coin flip means we were between 45% and 55% and the match settles nothing.

No gameweek has been played and scored yet, so there is no history. This is the honest empty state, not a missing feature — it fills in the first time scripts/score-gameweek.mjs runs after a completed gameweek.