The Numbers Underlying Revs Dramatic Turnaround
The Results Are On the Upswing -- But Do the Stats Agree?
A lot of times, when you start digging into soccer statistics, things don’t line up the way you might expect.
Sometimes the score-line says one thing while the stats indicate another. Much in the same way your eyes — while watching a game — might paint a different picture than the eventual result.
A team may look dominant but still lose big in the end. Or they could win while underperforming in every statistical category.
Usually those instances are the exception, not the rule. One-offs. Blips on the radar of a long season.
But what if those discrepancies aren’t a blip? What if they become a pattern?
For example, the Revolution appear to have righted the ship in recent weeks after a horrid start to their 2024 campaign. They’ve picked up 15 of their 22 points in just the past five weeks, racking up five wins in their last seven matches.
But there was something odd about those wins. While the box scores reflected wins, their expected goals differential (a statistical measure of the quality of chances each team generated) seemed to indicate that losses would have been the more fair result.
It was that discrepancy that got me wondering: Has New England actually turned a corner? Their results certainly seem to indicate yes. But, certainly, a sustainable shift in quality would see an accompanying uptick in statistical measures.
With that in mind, let’s dig into the stats and see how much, and in which ways, this team has changed as the season has worn on.
The Only Stat That Matters
Let’s lay the groundwork here, because I can already hear all the complaints about using convoluted advanced analytics to detract from actual goals and actual wins.
The Revolution have had very good results since the start of June. In the past seven games, they’ve won five and lost two. That’s good for 2.14 points per game, the 4th best pace in MLS over that timeframe.
They scored 11 goals over that seven-game span (1.57 goals/gm), picking up three wins at home and a pair of wins on the road, including at FC Cincinnati who currently hold the top spot in the league.
That surge in production was crucial for the Revs who — up until June 1st — found themselves with just two wins and one draw to show for 3+ months of league play. They entered June with a 2W-1D-10L record, having scored just nine times in those 13 games while conceding 26 times at the other end.
So, what changed in June that sent the Revs on this hot run of form?
Some Other Stats That May Matter
Honestly, less seems to have changed than you might expect given the turnaround in results.
The red bars in the graph above represent goals/game and expected goals/game during New England’s initial 13 games of the 2024 season, a period in which they managed just 0.54 points per game. The green bars show the same stats since June 1st, when the Revolution earned 2.14 points per game.
The first thing you may notice is an increase in Goals-For and a decrease in Goals-Against as we entered the summer months. That, however, does not coincide with a corresponding increase in expected Goals-For, nor a decrease in expected Goals-Against.
In fact, over the past seven games, New England is generating less xG and allowing more xGA than they did during their horrible 13-game opening stretch.
These are, of course, small sample sizes but even still it doesn’t match what you would expect of a team that had truly turned a corner.
Still, it’s clear that some things must have changed since match-day 1 back in February.
When we dig a bit deeper we see a few trends as time has gone on. Some of these are self evident, such as a decreased focus on possession.
With a high number of variables (scoreline, opponent, lineups, injuries, cards, etc.) the data was always going to look a bit noisy. There does, however, appear to be a loosely correlated negative trend in possession as the season has progressed. In the “bad-times” (aka the first 13 games) the Revs held an average of 54.2% possession while the “good-times” since then have seen them hold an average of 46.1%.
Hand-in-hand with the diminished possession is a decrease in the total number of passes as well.
Pass attempts are down about 30% from 557 per game to 388 per game.
In turn, so too did the number of shots decrease over time. Shots are down from an average of 12.85 to 11.57 per match.
This is only surprising given the fact that over the past month-ish the Revolution have turned those diminished shot attempts into goals at over double the rate they had before (0.69g/gm to 1.57g/gm).
With that increase in goals, surely the it would make sense that the quality of those shots improved? Or at the very least the shots were being taken from closer to goal?
Not quite.
Expected goals are down (though not by much) from 1.05xG/90 to 0.94xG/90. Additionally, the average shooting distance is up (perhaps a surprise for the “Chancalay shoots from too far away” crowd) from 17.71 yards to 19.57 yards per shot.
Wait, Huh?
Yeah fair enough. Let’s recap.
The Revs, since the beginning of June have:
Had less possession
Attempted fewer passes
Taken fewer shots
Generated fewer expected goals
Allowed more expected goals-against and…
Increased their average shooting distance
All while simultaneously scoring 2.3x more goals per match and winning in 5 out of 7 contests.
Sometimes the stats say one thing and the results say another.
I’m always interested to hear everyone’s theories (so drop a comment), but here are some candidate explanations.
The first, and most obvious, is the Revs seem to have leaned more into using direct attacks as the season has worn on. This means less time on the ball, fewer passes, and likely better-scoring scenarios. xG isn’t always a great reflection of just how good a scoring chance is, as even shots against fewer/less set defenders can have low xG values based upon where the shots are taken.
The Revolution have also scored first in each of those five wins, allowing them to take on a more defensive posture as the games wore on. This would explain the lower shot totals, as well as the higher xGA conceded to other teams.
Add in a spate of injuries on the attacking side of the ball and it’s not surprising that the Revolution haven’t generated a ton of high-quality looks.
It is, however, surprising that New England has been able to convert as many chances as they have. They scored 10 goals in their recent five wins, on just 4.5 expected goals.
It’s not uncommon to outperform your expected goals in games that you win (often times that’s how you win). But to produce more than double the goals you’d be expected to while also generating 1.0xG or less in three of those five wins is potentially concerning.
At the end of the day, the score is the most important number.
As long as the Revs continue to put together positive results these statistical trends won’t matter. But the underlying trends don’t show a team that has drastically improved from the team that won just twice through the first 3 months of the season.