Statistical Analysis: How many points is the attack costing Manchester United?

Manchester United are doing their best to throw away the Champions League.

They’re crawling to the finish line of the Premier League season, currently clinging to a top four spot by just one point. Recent form has turned a comfortable lead for the final place in the UEFA Champions League into a race much closer than anyone has comfort with.

There are several contributing factors into United’s recent poor form. One of those factors is the attack, which is the one we’re going to focus on today.

---

Manchester United needs a striker. Who that striker should be - or what type of player that striker should be - is a topic for another time but there is no argument that they need a striker. They need a striker because they do not score enough goals.

At the start of the season The Athletic surmised that teams looking to qualify for the Champions League should be looking to score at least 69 goals in the Premier League campaign. Manchester United have scored 49 goals through 34 games this season. That total is 9th most in the league - barely above average - and puts them on pace to score just under 55 goals this season, well short of the 69 you should be targeting.

Expected goals measure the probability of how likely every shot taken will result in a goal. It gives us a good barometer of how many goals United should have scored this year based on the chances they’ve created.

Per Opta, United have an xG of 55.1 goals this season, the sixth best total in the league. That isn’t great but teams with better players can be expected to overachieve against that number while teams lower down the table might be expected to underachieve on it. Taking the number of goals a team has scored and subtracting their xG can give us a good barometer of how good a team’s finishing has been.

That’s where United’s numbers are so astonishing. They’ve scored 8.5 fewer goals then their xG suggests they should have. That’s the fourth worst number in the league.

There’s no question that United’s finishing has been poor and it’s costing them goals. This has become even more prevalent after suffering back to back 1-0 defeats where United missed chances in the first half. The question is, how many points in the table is this costing them?

It’s easy to say that United should have eight more goals, look over the results of the campaign, divide those eight goals up over some low scoring matches and suddenly come away with a lot more points. That’s not exactly how it works though. You can’t just add goals in matches where it’s convenient. You have to look at which matches they actually under performed against the chances they created. Some of the under performance may have come in matches that United won.

In order to really answer this question, we have to look back at each game and look at whether they were over or underperforming against their metrics. We’ll take the number of (non-penalty) goals United scored in each game and subtract it by their non-penalty xG (United are 2/2 on penalties this season so we don’t need to worry about those) to see in which games there was under performance. Then we can add the goals United “should have” accordingly and see how that affects the score lines.

Let’s discuss the parameters.

What constitutes as underperforming and worthy as an extra goal?

There are 14 matches where United underperformed against their xG, but adjusting for underperformance would only affect the outcome of five of them. Nevertheless if they performed up to their metrics they’d have won an additional eight points this season, bringing their total to 71 points and putting them in a position where they’d need just one point over their last four matches to ensure Champions League qualification.

As you’re probably aware, this data is flawed. United’s npG-npxG is only -8.5 and there’s 14 goals in the table above. If we’re playing the ‘what if’ game, we have to play the full what if game. We can’t just cherry pick the things we’d like - United scoring more goals - and reward underperformance. If we’re doing that then we need to subtract for over performance as well.

So once again we have to look at the match by match npG-npxG numbers and add the goals in for underperformance but this time subtract goals for overperformance.

The parameters remained the same with one exception. We’re leaning towards scoring goals so I adjusted the line to remove a goal from +0.5 to +0.7, with the line to deduct a second goal still increasing by 1.0 (from 0.7 to 1.7).

Here’s how it looked.

If we take the sum of all the goal adjustments here the total is +8. Pretty damn close to the 8.5 Opta has. There were five matches where United drastically over performed on their metrics but none of the over performance really affected the final score. Even after adjusting for overperformance United still come away eight points better.

So can we say that United’s (anemic) attack has cost them eight points in the table?

It doesn’t seem like this data is telling us the full story. Again we’re only playing half of the what if game. What if United finished all their chances? What if they just won all their games??

It’s a far too simplistic way of looking at it because it doesn’t take into account that in each match United have an opponent and that opponent may have missed chances too!

It’s easy to say “oh if Rashford had just scored that header against Newcastle we would have won 1-0 and taken all three points.” Well, what if Joelinton had finished his header earlier in the match? Then Newcastle could have won 1-0 and taken all three points! (Both teams were denied penalty shouts in that match as well). West Ham scored from a terrible shot on Sunday. It’s easy to say that cost United a point, but what if West Ham had just finished one of their other chances?

As such, let’s look at the same under/overperformance line of all of United’s opponents this season.

And now let’s look at how that over or underperformance would affect the scoreline of each match, combined with United’s over or underperformance.

Again we see most of the outcomes wouldn’t have been affected, but while better finishing from United would have created more favorable results in four matches, better finishing from their opponents would have created worse results in three others. The end result is a net gain of two points. Not much, and while every point matters at this stage of the season, it wouldn’t really change what United need to do over their final four games.

(There were also two results that I didn’t adjust. Brentford (H) had an underperformance of -0.7 and should have gotten a goal while Aston Villa (H) were -0.6. I didn’t adjust these matches because the expected goal difference for each match came out to just over 0.5 and I deemed a half a goal difference as enough to say United still should have come out on top. Had we included those results, the cumulative points adjustment would be -2).

It’s been a weird season. There’s no question that United need to score more goals but it’s hard to say that’s what cost United this season. United have played 15 one goal games this season, but they’ve won 11 of them! They’ve only drawn six games but only in two of them can they say they underperformed more than their opponent did.

For most of the season United have been playing on very thin margins but still coming out on top. In the last two weeks, they played on the same thin margins only this time they came out on the wrong side. You can argue they got off to better starts in both games, and finishing early chances would have changed the game state of each game. But it could also be pointed out that just a few weeks ago United scored six minutes into the game and took a 2-0 lead into halftime against Tottenham only to leave with a 2-2 draw.

So while United needed to score more goals, those goals wouldn’t have helped them too much because they conceded too many goals.

Five times this season United have conceded three or more goals. In most of those games a goal or two wouldn’t have changed the result. If you wanted those goals to matter you’d have to concede less.

United have conceded 40 goals this season - only the seventh fewest total in the league. They’ve gotten this far because over half of that total (23) have come in just five matches (Brentford (A), City (A), Aston Villa (A), Arsenal (A), Liverpool (A)). Conceding all the goals at once means that in all the other games the goals United score go much further.

They’re also in this position because their opponents miss chances too. United’s non-penalty expected goals against of 43.7 is 9th best in the league. As close to league average as you’ll get, and their npG-npxG comes in at -5.7! Only five teams have seen their opponents underperform more but in the case of those five, it’s thanks to superior play from their goalkeeper with all five teams goalkeepers saving more shots than would be expected. David de Gea has conceded 1.7 more goals than would be expected based on the shots he’s faced.

To say United just need to do a better job finishing their chances would be a gross miscalculation of what this team needs. Yes they need to score more goals but in order to do that they need to create more chances. This team is currently fifth in shots per game, sixth in non-penalty xG, and seventh in xG per shot, which tells us they’re not creating high quality chances either.

How do you improve that? Well looking at the data MarkStats collects shows us exactly where the problem is. Per MarkStats United are currently sixth in average possession, seventh in field tilts, and seventh in expected threat. They don’t have the ball enough, and when they have it they struggle to effectively move it into dangerous positions.

Marcus Rashford has the highest conversion percent and npG-npxG of his career. Conversely, Antony has the lowest of his. It’s not unreasonable to expect that to turn around next season., but in order to really take advantage of that, they’re going to need to create more chances.

That’s the crux of matter. Sure they need to be more ruthless in front of the net, but oddly, their lack of finishing really hasn’t cost them much this season.