Matchday 2 is the first round of the season this division carries model probabilities at all. The 1X2 model was silent through Matchday 1 because it needs a completed match per side before it will forecast anyone; there were no scoreline picks either, for a separate reason — three of the fourteen clubs came up this summer holding no rating at all, which left too few rated fixtures to work across. Seven matches over two days: one at 12.45pm on Saturday, then one at 12pm, three at 2pm, one at 2.45pm and one at 6.30pm on Sunday. Read the numbers knowing how thin the evidence under them still is: one match played, and for the three promoted clubs most of the form window is last season’s divisional average rather than anything they have done.
Gameweek 2 – Stats & Historical Data
All Fixtures — Head to Head, Form & Historical Positions
Wsl - Matchday 2 Preview
Season 2026-27 | Powered by 6 seasons of historical data
West Ham W10th
London City Lionesses3rdxGF: 4.8 | xGA: 3.7 (+1.1)
xGF: 8.0 | xGA: 4.6 (+3.4)
12 Sep 2026 - 2026-27
Manchester United W5th
Chelsea W1stxGF: 4.0 | xGA: 3.5 (+0.6)
xGF: 11.0 | xGA: 3.0 (+8.0)
13 Sep 2026 - 2026-27
Birmingham City W12th
Brighton W6thxGF: 4.3 | xGA: 6.0 (-1.8)
xGF: 5.3 | xGA: 4.0 (+1.2)
13 Sep 2026 - 2026-27
Liverpool W11th
Tottenham Hotspur W4thxGF: 3.4 | xGA: 4.6 (-1.3)
xGF: 4.4 | xGA: 5.7 (-1.4)
13 Sep 2026 - 2026-27
Everton W8th
Charlton Athletic W14thxGF: 1.8 | xGA: 4.2 (-2.3)
xGF: 1.3 | xGA: 9.5 (-8.2)
13 Sep 2026 - 2026-27
Arsenal W7th
Crystal Palace W13thxGF: 7.8 | xGA: 1.3 (+6.4)
xGF: 2.7 | xGA: 7.4 (-4.7)
13 Sep 2026 - 2026-27
Aston Villa W9th
Manchester City W2ndxGF: 2.8 | xGA: 4.6 (-1.8)
xGF: 6.0 | xGA: 5.3 (+0.7)
13 Sep 2026 - 2026-27
Matchday 2 - Historical Insights
Playoff Race: Teams in 6th place at matchday 2 have averaged 3 points historically.
Current WSL Table
WSL Table
Women's Super League
Updated: 03 Nov 2026, 12:09 PM
Key Storylines This Week
What to Watch For
- The first round the model has numbers for: Matchday 1 carried no probabilities and no scoreline picks; everything on the model tab is the first of its kind this season, drawn from a single round of evidence.
- Three clubs have yet to concede, three have yet to score: Arsenal W, Crystal Palace W and Liverpool W kept opening-weekend clean sheets; Brighton W, Charlton Athletic W and Everton W have not scored. After one match neither list means much, but they are close to the only separation there is — every club has played once and most of the table is ties.
- Two models, and three fixtures they read differently: The bars and the scorelines under each fixture come from two different models. The model calls five home wins, zero draws and two away wins across the seven fixtures, and the closest score consistent with each call breaks down by margin as six by one, one by two. The goal rates behind them run from 0.72 to 2.61 a side. On three of the seven they name different winners, and each of those fixtures says so beneath it.
WSL Top Scorers
Top Scorers Heading into Gameweek 2
Top Assists — 2025/26 season
Last season’s assist leaders, shown for context. The club against each name is the one they played for in 2025/26; some will have moved over the summer.
Top Assist Providers
WSL Top Assists
Head to Head
Manchester United W vs Chelsea W
Last 5 league meetings
| Date | Home | Score | Away | xG |
|---|---|---|---|---|
| 13 Sep 2026 2026-27 | Manchester United W | 0 - 5 | Chelsea W | - |
| 16 May 2026 2025-26 | Chelsea W | 1 - 0 | Manchester United W | - |
| 03 Oct 2025 2025-26 | Manchester United W | 1 - 1 | Chelsea W | - |
| 30 Apr 2025 2024-25 | Manchester United W | 0 - 1 | Chelsea W | - |
| 24 Nov 2024 2024-25 | Chelsea W | 1 - 0 | Manchester United W | - |
Recent Form Comparison — end of 2025/26
Form Comparison - Last 6 Games
| # | Team | P | W | D | L | GF | GA | GD | Pts | Form |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Chelsea W | 6 | 5 | 1 | 0 | 15 | 7 | +8 | 16 | WDWWWW |
| 2 | Manchester United W | 6 | 1 | 3 | 2 | 3 | 6 | -3 | 6 | DWLDDL |
Manchester United W - Recent Results
Chelsea W - Recent Results
Last Season at a Glance
Manchester United W
WSL 2025/2026 Season Summary
Overall
Home
Away
xG Analysis
Based on 10 matches with xG data
Overperforming xG by 38.0 goals - clinical finishing or some luck involved.
Chelsea W
WSL 2025/2026 Season Summary
Overall
Home
Away
xG Analysis
Based on 9 matches with xG data
Overperforming xG by 44.0 goals - clinical finishing or some luck involved.
Goals by Period — 2025/26 season
Manchester United W - Goals by Time Period 2025/2026
Chelsea W - Goals by Time Period 2025/2026
Historical Position at Matchday 2
Historical Position Comparison
Manchester United W - Sunday 13th September 2026
| Season | Pos | Pts | GD | Final |
|---|---|---|---|---|
| 2026-27 (now) | 5th | 12 | -2 | - |
| 2025-26 | 1st | 6 | +8 | 3rd |
| 2024-25 | 2nd | 6 | +4 | 3rd |
| 2023-24 | 4th | 4 | +1 | 5th |
| 2022-23 | 2nd | 6 | +5 | 2nd |
| 2021-22 | 3rd | 6 | +4 | 4th |
| 2020-21 | 4th | 7 | +6 | 4th |
| 2019-20 | 12th | 0 | -2 | 4th |
Historical Position Comparison
Chelsea W - Sunday 13th September 2026
| Season | Pos | Pts | GD | Final |
|---|---|---|---|---|
| 2026-27 (now) | 1st | 22 | +23 | - |
| 2025-26 | 3rd | 6 | +3 | 2nd |
| 2024-25 | 1st | 6 | +8 | 1st |
| 2023-24 | 5th | 4 | +1 | 1st |
| 2022-23 | 5th | 3 | +1 | 1st |
| 2021-22 | 6th | 3 | +3 | 1st |
| 2020-21 | 3rd | 7 | +10 | 1st |
| 2019-20 | 4th | 4 | +1 | 2nd |
Player to Watch — 2025/26 season
Lauren James — Chelsea W
Lauren James scored one for Chelsea W on the opening weekend, as many as anyone across the two clubs. One match is not form, and with no expected-goals data in this division there is nothing underneath the number to say whether the chances were good ones. The radar below is last season’s data.
Lauren James
Chelsea W • M • 2025-2026 Season
BTP Verdict
The margin is the number worth reading rather than the score: the model puts 47% on this finishing within a goal, the second-lowest of the seven — it expects this one to spread further than most. The likeliest scorelines are 0-2 12%, 0-1 11%, 1-2 9%, every one of them a Chelsea W win, which is the goals model disagreeing with the bars above in plain sight.
The two models disagree about three of the seven
Every fixture on this page carries two separate model outputs: the probability bars come from the 1X2 classifier, and the scorelines beneath them come from a goals model that estimates a scoring rate for each side. They usually point the same way. This week they do not, on three of the seven — Manchester United W v Chelsea W, where the bars lean Manchester United W but the goal rates make Chelsea W 74%; Everton W v Charlton Athletic W, where the bars lean Everton W but the goal rates make Charlton Athletic W 42%; Liverpool W v Tottenham Hotspur W, where the bars lean Liverpool W but the goal rates make Tottenham Hotspur W 41%. We publish both rather than quietly pick the one that reads better. A fixture the two disagree about is one neither of them is confident in, and after a single matchday that is the honest position to be in.
What the model does not know about three of these clubs
The goals model works from each club’s last five matches in this division. Birmingham City W, Charlton Athletic W and Crystal Palace W came up this summer, so four of those five slots are filled with last season’s divisional average instead — which means the grid under three of this week’s fixtures is largely a league baseline wearing a club’s name. Each of those fixtures says so beneath it. It is an absence of data rather than a low rating, and no import will fix it: the provider publishes no player or match statistics for the division they came from.
One heavy defeat, and a round that was otherwise close
The opening round produced nineteen goals from seven matches, split four home wins, one draw and two away wins. Four of the seven finished inside a single goal, and only one was decided by more than two — Liverpool won 4-0 at Charlton, in Charlton’s first match at this level. Crystal Palace W were the only one of the three promoted clubs to win, beating Everton 1-0 at home — one of the four home wins above. One round is far too little to read a season from, and at twenty-six matchdays there is a great deal of it left.
Every draw figure here is about half the Championship’s
The draw probabilities across these seven run from 10.8% to 14.5%. In the Championship’s round this weekend the same figures run from 20.3% to 26.5%, roughly double. Part of that is real: across 2019/20 to 2025/26 this division drew 17.3% of its matches against the Championship’s 26.0%, so fewer draws is a property of the football and not an opinion of the model’s. But only part — the model averages 12.4% here, below even the WSL’s own long-run rate, which is the same three-way-split arithmetic that stops it ever calling a draw outright. Treat a draw bar in this division as meaning more than its size suggests. And note the harder limit while we are on what the numbers cannot do: the WSL has no expected-goals data in any season, so nothing on this page separates a good chance from a poor one.
Player Watchlist
Three players from three different clubs to keep an eye on this weekend.
- Grace Fisk (Liverpool W) — Two goals in one match for Liverpool W, two of the four they have scored.
- Khadija Shaw (Manchester City W) — Two goals in one match for Manchester City W, two of the three they have scored.
- Kirsty Hanson (Tottenham Hotspur W) — Two goals in one match for Tottenham Hotspur W, two of the three they have scored.
BTP Predictions — Gameweek 2
Final Thoughts
Seven matches, the first probabilities of the season, and a model that is honest about being early: it disagrees with itself on three fixtures and leans on a league average in three others — Everton W v Charlton Athletic W is in both groups. Manchester United W v Chelsea W is the pick of the round, and the fixture its two halves read most differently.
West Ham W vs London City Lionesses
📊 ML Probability
Match Prediction
The model’s most likely outcome is London City Lionesses, at 33.0% / 10.8% / 56.2% home, draw and away. It puts 62% on the match finishing within a goal; the likeliest scorelines are 0-1 11%, 1-1 10%, 1-2 10%, none of them close to certain.
Manchester United W vs Chelsea W
📊 ML Probability
Match Prediction
The model’s most likely outcome is Manchester United W, at 46.2% / 12.2% / 41.6% home, draw and away. It puts 47% on the match finishing within a goal; the likeliest scorelines are 0-2 12%, 0-1 11%, 1-2 9%, none of them close to certain. The two models part company here: the bars come from the 1X2 model, while the goal rates behind those scorelines read it 12% / 15% / 74% home, draw and away, favouring Chelsea W. We publish both rather than pick between them — a fixture the two disagree about is one neither is confident in.
Birmingham City W vs Brighton W
📊 ML Probability
Match Prediction
The model’s most likely outcome is Birmingham City W, at 60.5% / 11.8% / 27.7% home, draw and away. It puts 65% on the match finishing within a goal; the likeliest scorelines are 1-1 10%, 1-0 10%, 0-1 9%, none of them close to certain. One caveat on the grid: Four of the five matches behind Birmingham City W’s side of it are last season’s divisional average rather than their own results, because Birmingham City W last played a season in this division in 2021/22. Treat those figures as the division’s baseline adjusted for the little that is known, not as a reading of the club.
Everton W vs Charlton Athletic W
📊 ML Probability
Match Prediction
The model’s most likely outcome is Everton W, at 65.4% / 12.2% / 22.4% home, draw and away. It puts 65% on the match finishing within a goal; the likeliest scorelines are 1-1 10%, 0-1 10%, 1-0 9%, none of them close to certain. The two models part company here: the bars come from the 1X2 model, while the goal rates behind those scorelines read it 36% / 22% / 42% home, draw and away, favouring Charlton Athletic W. As above, both readings are published rather than one chosen. One caveat on the grid: Four of the five matches behind Charlton Athletic W’s side of it are last season’s divisional average rather than their own results, because there are no matches for Charlton Athletic W in this division across the seasons the model draws on. Treat those figures as the division’s baseline adjusted for the little that is known, not as a reading of the club.
Liverpool W vs Tottenham Hotspur W
📊 ML Probability
Match Prediction
The model’s most likely outcome is Liverpool W, at 52.2% / 14.5% / 33.3% home, draw and away. It puts 64% on the match finishing within a goal; the likeliest scorelines are 1-1 10%, 0-1 10%, 1-0 9%, none of them close to certain. The two models part company here: the bars come from the 1X2 model, while the goal rates behind those scorelines read it 37% / 22% / 41% home, draw and away, favouring Tottenham Hotspur W. As above, both readings are published rather than one chosen.
Arsenal W vs Crystal Palace W
📊 ML Probability
Match Prediction
The model’s most likely outcome is Arsenal W, at 53.4% / 13.0% / 33.6% home, draw and away. It puts 55% on the match finishing within a goal; the likeliest scorelines are 1-0 11%, 2-1 10%, 2-0 10%, none of them close to certain. One caveat on the grid: Four of the five matches behind Crystal Palace W’s side of it are last season’s divisional average rather than their own results, because Crystal Palace W last played a season in this division in 2024/25. Treat those figures as the division’s baseline adjusted for the little that is known, not as a reading of the club.
Aston Villa W vs Manchester City W
📊 ML Probability
Match Prediction
The model’s most likely outcome is Manchester City W, at 38.5% / 12.4% / 49.1% home, draw and away. It puts 42% on the match finishing within a goal; the likeliest scorelines are 0-2 12%, 0-3 11%, 0-1 10%, none of them close to certain.
🤖 Model Notes — GW2
Two different things appear on this tab and they are worth separating. The probability bars are our Women’s Super League model (wsl_lr_v2_cal), which has produced rows for every fixture from Matchday 2 onwards — there were none at Matchday 1, because it needs at least one completed match per side.
The scorelines come from a separate model: a Poisson goals model with a Dixon-Coles low-score correction (wsl_poisson_v1). It estimates a goal rate for each side from their last five matches in this division — reaching back into last season where a club has them, not this season alone — then builds a grid of every plausible score.
What the table shows for each fixture is the margin — how likely the match is to finish within a goal — and then the three likeliest exact scores with their probabilities, taken straight from that grid. They are listed outright rather than filtered to agree with the call above them, so on a fixture where the two models pull against each other you will see scorelines that do not match the bars. That is the grid being shown as it is, not an error. A single pick appears in only one place, the match of the week, where it is the conditional modal scoreline: the likeliest exact score consistent with the 1X2 call, which is why the percentage beside it is lower than the top line of the table.
Where a club has played fewer than five matches in this division, the missing games are filled with last season’s divisional average rather than left blank. We say so beneath a fixture when a side is leaning on that filler for three or more of its five, and publish no scoreline at all where both sides are leaning on it for all five, because the model would then return the same figures it would give any two clubs in that position. The most any side leans on it this week is four of the five, and three fixtures are at or past the three-slot threshold, so they each carry a note beneath the grid saying which side it applies to.
There is a probability chart below for each of the 7 fixtures being previewed, not a selection of them.
Model: Platt-calibrated Logistic Regression with position-segmented, z-scored player features (wsl_lr_v2_cal, deployed 27 May 2026). Trained on WSL seasons 2020/21–2024/25. Supersedes wsl_lr_v1_cal — v2 walk-forward log-loss on 2024/25 holdout is ~6.6% lower than v1’s 0.9683. Companion Poisson + Dixon-Coles goals model (wsl_poisson_v1, deployed 5 Aug 2026, ρ=+0.084) drives scoreline grids: 25/26 walk-forward log-loss 0.909 vs naive 1.040 (+12.6%). Full methodology here.
Check back after the weekend for our full Gameweek 2 WSL review with match reports and updated standings!
