Have Blades dropped off, or is the data just catching up?

Tom Hawkins compares the contrasting data between Blades, Burnley and Leeds as the automatic promotion battle seems to have slipped from our grip.

Tom Hawkins

The battle for automatic promotion to the Premier League is heating up, with new plot twists seemingly emerging after every match. This past week, it’s been Sheffield United’s turn to stumble, falling to 1-0 defeats against both Oxford United and Millwall, before an even more embarrassing 2-1 loss away at bottom-of-the-table Plymouth Argyle. There’s been plenty of grumbling among the Blades fanbase after these disappointing results — and understandably so — with concerns about our form dipping so suddenly at the business end of the season.

But has our performance level really dropped that much in these last three games, or is it just that the data has finally caught up with us?

Are we still in a good position to catch Leeds United and Burnley at the top, or have we already blown our chance by surrendering nine points in the last three matches?

Well, let’s dive into the data — the same data that Wilder has previously, politely, told us where to stick. First off, take Leeds United, who currently sit top of the table. They have a goal difference of +53, which actually overperforms their expected goal difference (xGD) of +49.3 (as per FBref) — which surprised me, given their goalkeeper Meslier seems to want to chuck one in his own net every couple of games. Despite that overperformance, it still equates to an xGD of +1.17 per 90 — meaning, on average, they’re creating over a goal’s worth more chances than their opponents in every match.

That’s quite a staggering difference compared to the xGD per 90 of Sheffield United (+0.32) and Burnley (+0.33). In this case, xGD seems to have mattered a little less over the course of the season, thanks to the frankly ridiculous overperformance of goalkeepers Michael Cooper and James Trafford, who have prevented 4.93 and 12.5(!) goals respectively based on the xG on target (xGOT) they’ve faced (according to FotMob).

If we look at the graph below, we can see the five-game rolling mean xG ratio of each of these teams throughout the season so far. This is meant to capture short-term performance levels, by looking at xG created and conceded over the last five matches.

In terms of interpreting the graph: red lines represent data for Sheffield United, with yellow and blue lines corresponding to Leeds and Burnley respectively. The bold lines show the five-game rolling xG ratio, while the faded dotted/dashed lines in the background provide further detail about the average xG being created and conceded during that stretch.

The graph really highlights the dominance of Leeds’ performances over the season — including a particularly impressive period where they were averaging over 2 xG more than their opponents. For Sheffield United and Burnley, it’s been a different story: more staggered performance levels throughout the season and far less dominance. That includes periods where both teams were consistently creating less xG than their opponents (watch for where the bold lines drop below 0).

So why is it, then, that despite Leeds’ apparent dominance throughout the season, all three teams have remained neck-and-neck at the top — especially if we conveniently ignore the two-point deduction for Sheffield United? It’s easy to lean on the classic “Leeds are falling apart” narrative, which explains why they aren’t further ahead, but the truth might be a little more nuanced than that.

For one, as mentioned earlier, both Sheffield United and Burnley have benefitted from some truly outstanding goalkeeping performances this season — and personally, I believe that if Leeds had either of those keepers instead of Meslier, they’d probably have comfortably wrapped up the title by now.

But there’s also something to be said for the idea of game state and how it affects xG created and conceded. Game state essentially describes the current scoreline of a match — whether you’re winning, drawing, or losing. This season, Sheffield United have been particularly good at taking slender leads and seeing out results by shutting up shop. That approach effectively invites the opposition to come forward and create more chances, while we don’t go chasing more goals ourselves. This likely skews our xG figures throughout the season.

Leeds, on the other hand, tend to seek total control of the game and continue pushing to add more goals — which perhaps leaves them with a softer underbelly, more vulnerable to conceding the occasional cheap one.

So where does that leave us, with four games left to play? From the graph, it’s clear that all three teams are finding it tough to maintain consistent performance levels right now, with nerves potentially denting confidence as the finish line approaches. Obviously, this is having a bigger impact on some than others.

Sheffield United’s main issue has been struggling to chase games when going behind — as seen against Oxford United and Millwall — because we’re usually so reliant on grabbing that all-important first goal. We also looked very lacklustre when trying to see out what should have been a straightforward win against Plymouth Argyle.

The truth is, none of these games have been wildly different from other similar performances earlier in the season, where we managed to scrape wins with a heavy slice of luck on our side. So maybe these recent results are just a case of us regressing towards what we’ve actually deserved from these kinds of performances all along.

Although it feels bleak now, there’s no reason that the race for automatic promotion can’t go right down to the wire. Five points feels like a bit of a mountain to climb with four games remaining, but if we can replicate the performance levels we showed in our win over Coventry City — and get a crucial result against Burnley on Easter Monday — the complexion of the promotion race could change very quickly.