The Worryingly Predictable Nature of the Women's Super League
Surprise results have been at a premium in the 2024/25 Women's Super League. That doesn't mean a different team shouldn't be top according to the data.
The Women’s Super League returns from its winter break this weekend. As we are essentially half way through the 2024/25 season, with 10 of the 22 match weeks completed, it feels a good time to take stock of how the table is shaping up.
One aspect of what makes sport so exciting is its unpredictability. We are frequently told that a selling point of the Premier League is that any team can beat any other on a given day.
Upsets are less frequent than they would’ve been in the old days thanks to the financial disparity in the division. Even so, Arsenal are the only one of the established big six clubs yet to lose a home game which bookmakers expected them to win, with their two stumbles by this measure occurring on the road (at Bournemouth and Newcastle).
Across the first 199 Premier League matches of 2024/25, there were 42 which were won by the bookies’ underdog. This gives us an upset rate of 21 per cent. How often do you think there is a similar surprise result in the WSL?
The only one which probably immediately springs to mind is Everton’s 2-1 win over Manchester City in the last round of fixtures. The Blues were rated as having just a seven per cent chance of claiming the three points yet managed to do so. There have only been five other upsets, with two of them seeing the favourite having no more than 4.1 per cent more chance of winning than the underdog anyway. Even with these very mildly unexpected results, the upset rate in the WSL this term is only 10 per cent.
A couple of things stood out from these matches. Liverpool’s win at Tottenham needed a 95th minute penalty, else it would’ve been a draw. And while newly promoted Crystal Palace winning away anywhere will be unexpected this season, they did so against the team currently only one point better off than them, Leicester.
Digging deeper after noticing that revealed a trend across the division. The Foxes got a surprise draw against Chelsea but their only victory occurred against fellow strugglers Everton. Tenth placed West Ham have an even starker set of results; wins over the two teams below them, draws against the two directly above and defeats against the top seven sides.
On the Wikipedia page for the 2024/25 WSL season there is a grid of the results. It lists the teams in alphabetical order, with blue denoting home wins, yellow used for draws and red indicating away victories.
Rearranging the grid into league position order confirms everything written above. Upsets, even among fairly closely matched teams, are a rarity.
There are no easy solutions to this issue. A look at the underlying data suggests there have been more surprising - or perhaps, undeserved - results than we might think though.
There have been studies in the past (e.g. by Statsbomb) about whether expected goals should be the same for women’s football as it is for men’s. Using WSL data from FBRef, I treated it identically in the sense of saying that if one team has more than 0.4 xG more than they opposition then they deserve to win. As the below table shows, perhaps I shouldn’t have.
This method suggests that Arsenal have deserved to win every match this season. Their narrowest expected goal margin has been 0.6 in their favour, in their goalless draw with Everton. Not until you raise the bar of xG to 1.0 to determine a merited win do they fall from at worst joint-top of the expected WSL table.
Across the season as a whole, the Gunners are not miles off their expected figures at either end. Yet that in itself might be a problem at the top end of women’s football; Arsenal have been one per cent below their xG up front and 14 per cent better off at the back, but those figures for Chelsea are +37 and +50 respectively. The champions are breaking the model.
The Blues have also only missed 17 of their 36 Opta-defined big chances whereas their north London rivals have squandered 27 of 38. Manchester City have only converted 12 of their 32, suggesting this old-school metric alone may explain much of the reason why Chelsea are top.
The WSL table shows that Manchester United should also be assessed among the title contenders. Their credentials are undermined by the expected points standings showing that only Brighton have earned more points than they theoretically deserved.
United manager Marc Skinner owes a debt of gratitude to goalkeeper Phallon Tullis-Joyce. She has saved 6.1 goals more than expected, putting her on track to record the biggest over performance seen across the seven WSL seasons for which advanced data is available. Sorry, Mary who?
Problems at this end of the field have contributed to last season’s fourth placed team Liverpool finding themselves in eighth, though over a quarter of the goals they have conceded being deflected or put through their own net hasn’t helped. Only Tottenham (-5.9) have performed worse against their post-shot expected goals conceded than the Reds have (-3.3) , with neither Teagan Micah (-1.8) or Rachael Laws (-1.5) able to solve the issue between the posts.
Liverpool’s underlying performances as a collective haven’t been wildly different from last season. Against the corresponding fixtures, an xGD drop of -2.1 has cost them seven points and nine in the expected goal difference column.
However, this shortfall can be assigned to just two fixtures. Last season, Matt Beard’s side beat Chelsea at home and United away, with their losses in those matches this season seeing a negative goal difference swing of nine, plus three in xG terms.
They would be having an almost identical campaign to 2023/24 if those results were struck from their record. But then based on the pre-match odds they were two of the six biggest upsets in the WSL last season, to bring us full circle. The largest of the lot is unlikely to be repeated based on the expected goal performance data for this season either: Liverpool’s 1-0 win at the Emirates Stadium.
It is pointless to expect unlikely results like those to reoccur too often. Without such upsets, the WSL is looking a little too predictable for its own good though.
I had a ticket to watch Liverpool Women yesterday, so planned to write something on them, but the match was postponed. Still, the half way (ish) point of the season is a good time to assess things.