How "lucky" were Manchester United last year vs how good they actually were?
The 2020-21 version of Manchester United were known for constantly coming from behind to win games, a pattern often labelled as lucky and unsustainable. But how "lucky" were they really?
Ask 100 Manchester United fans what’s gone wrong this season and you’re probably going to get close to 100 different answers.
A common answer that I’ve seen in various places - whether it be Twitter, Facebook, a comment section, overheard at the bar etc - has simply been, “their luck ran out.” “Last season they were coming from behind and getting lucky to win so many games. There were so many games where if the things just went differently they wouldn’t have won. This year the luck has run out and we’re not getting those comebacks.”
Ok, I can’t argue with “so many games where if things went differently the result could have been different.” In fact, I’ll go as far as to say that United played 38 Premier League matches last season - and another 20 this season - where if things on the pitch had played out differently the result could have been different.
The more removed we get from last season, the more there seems to be a big(ger) discrepancy between how lucky United were last season and how good they were.
A lot of that I believe comes from the word “comeback.” In football, it’s used far too liberally. United came back to win a lot of games last season, it’s unsustainable to do think they can do that again. Not exactly, there’s a lot of context missing from that statement.
As soon as someone goes one nil down in football we start talking about ‘came from behind, which is a little ridiculous because not every comeback carries equal weight. For starters time on the clock matters, a lot.
A few weeks ago Liverpool went 2-0 up in the first half against Chelsea. Chelsea leveled the score just before half time prompting the announcer to say “Chelsea have completed this magnificent comeback.” Let’s not get ahead of ourselves. Personally I don’t find any comeback that happens in the first half to be “magnificent” for the sole reason of “there’s more time left on the clock to score the same amount of goals then the time it took your opponent to score them.”
Therefore the more time that has gone by, the more impressive the comeback. Being 1-0 down in the 90th minute and winning 2-1, that’s incredible. Being 3-1 down in the 70th minute and winning 4-3? Also incredible. Being 2-0 down after 15 minutes then being the better team over the next 75 minutes and winning 3-2? A bit less so. Being 1-0 down after 20 minutes, leveling before halftime and winning 2-1? We’re still calling that a comeback?
If United were scoring 85th minute equalizers and injury time winners week after week1 then yes, it’d be fair to say they were getting lucky and this was unsustainable. But rarely did they leave things so late. It’s still football so there’s still an element of luck. There’s so much variance the better team doesn’t always win, sometimes they just get unlucky on a day. Sometimes they get lucky and find a winner, but sometimes their opponent were the ones getting lucky when they scored the first goal to even get the lead!
Yes goals change the game state and that will play an effect on how things play out from there but how “lucky” is a late winner really? Is an 88th minute winner in a 0-0 match less “lucky” than an 88th minute winner in a match where you scored an equalizer in the 63rd minute? Or is it simply that late goals are *most of the time* just scored by the better team who have been knocking on the door most of the game and finally got their breakthrough?
If a game is 1-1 after 60 minutes and one team dominates the entire second half en route to a 3-1 win, does it really matter what order the first two goals were scored in? It’s 2022, we have numbers that can tell us a lot about a match and whether a team was a deserved winner or not.
As for the 2020-21 version of Manchester United. How lucky were they? According to FBref (via Statsbomb) United were fourth last season in xG difference at +18. Next on the list was Brighton (+13.9) followed by Arsenal (+9.2). Were they lucky to finish second? Certainly, but this was still firmly a Champions League caliber team.
Understat’s expected points model2 shows United being a little luckier. United finished with 77 points compared to an expected points total 65.64. That's a difference of 8.36 points but still firmly the fourth best team in the league. Yes, that’s a pretty big discrepancy but you expect teams with good players to overachieve a bit while teams without good players will typically underachieve a bit.3 A year before United actually underachieved, finishing with 66 points compared to an xPTS total of 70.99. That's 4.99 fewer than expected but again still firmly the fourth best team in the league. Their expected goal differential per Statsbomb was also fourth best in the league, firmly ahead of fifth place Leicester City. Luck or not, United were one of the four best teams in the league the past two seasons.4
But they made a lot of comebacks! That’s unsustainable and “lucky.”
Well how lucky did they really get? In how many of those games did they get lucky, or were they often just the better team that should have won anyway?
Let’s take a look5 at all the games where United came from behind as well as what the expected goals were for each game. We’ll look at what actually happened, try to factor in game state, and then gage how lucky they were on a scale of 1-10 (10 being the luckiest). Is this a perfect system? Of course not, because ultimately that number is just one man’s opinion.
(Disclaimer: We’re gonna use both Fbref (via Statsbomb) xG - because it’s better - as well as understat xG - because we can get the xG values of each shot - for each match)
Brighton 2-3 Manchester United
Statsbomb xG: 2.4 - 1.1
Understat xG: 2.98 - 1.58
What Happened? United went down after Bruno gave away a soft penalty, only to equalize three minutes later. They had a couple second half goals called back for offside but United were cruising to a deserved win before they went looking to bring trouble upon themselves.
The non-penalty xG was 0.77 - 0.39 in United’s favor at the 85th minute mark when Ole Gunnar Solskjaer oddly decided to switch to a back five and invited tons of pressure. Brighton racked up 1.15 xG over the next 5+ minutes to get a deserved equalizer only for United to somehow get a penalty after time expired.
Luck rating: 10 - you don’t get luckier than a penalty after the final whistle
Newcastle 1-4 Manchester United
Statsbomb xG: 0.5 - 2.7
Understat xG: 0.87 - 2.22
What Happened? Newcastle scored from an own goal three minutes in and hardly threatened again. United were by far the better team over the next 87 minutes and equalized before half time (despite missing a penalty). A weird one only in the sense that the two goals in injury time actually made the scoreline more indicative of how the match really looked.
Luck rating: 2 - United were thorough winners and it’s only a mystery how it took them so long to score a second
Everton 1-3 Manchester United
Statsbomb xG: 0.6 - 1.6
Understat xG: 0.38 - 1.60
What Happened? Bernardo scored a screamer completely against the run of play to open the scoring. 13 minutes later United were 2-1 up.
Luck rating: 1 - United were by far the better team over the full 90 minutes.
Southampton 2-3 Manchester United
Statsbomb xG: 0.4 - 2.3
Understat xG: 0.50 - 2.48
What Happened? By far the weirdest one on the list because it’s rare to see the numbers correct themselves so thoroughly over the course of just a single game. United didn’t do much in the first half in attack but held the Saints to just 0.23 xG. Yet they found themselves undeservedly 2-0 down at halftime thanks to two goals from set pieces.6 United pounded the Saints in the second half. They equalized with just under 20 minutes to play before Edinson Cavani’s 91st minute winner.
Luck rating: 3 - it appears more lucky then it was thanks to the dramatic fashion it was won in. Play that match over and over again and United would win far more often than not, just in less dramatic fashion.
West Ham 1-3 Manchester United
Statsbomb xG: 2.2 - 1.9
Understat xG: 2.53 - 1.79
What Happened? BIG time the numbers don’t tell the story game. West Ham pounded United in one of the worst 45 minutes United played all season but somehow only went in to halftime up 1-0. It can’t be understated how lucky United were to only be down 1-0 here. The second half was better but it still took a dodgy goal that probably shouldn’t have counted to level things up. That changed the game state which opened up space in behind for United to attack and it was all United from there. United were by far the better team in the second half - and Rashford even missed a sitter that could have given them four, but United should never have even been in a position where a comeback was possible this game.
Luck rating: 9.99 - the only reason this isn’t a 10 is because it’s not on the same level as “won a penalty after the full time whistle”
Sheffield United 2-3 Manchester United
Statsbomb xG: 1.6 - 2.0
Understat xG: 1.24 - 1.79
What Happened? A horrific error from Dean Henderson saw United go behind with only 84 minutes left to play (41% of SHU xG came from that chance). Just after the half hour mark United were 2-1 up and added a third five minutes into the second half. Then it was cruise control and by the 70th minute they were just going through the motions - though they lost some control once Bruno Fernandes came off. A very unlucky late goal from a corner where Victor Lindelof won the header but had it deflect off of David McGoldrick and into the net made this match seem a lot dicier than it actually was.
Luck rating: 2 - United were in control the whole way barring one really bad individual mistake
Fulham 1-2 Manchester United
Statsbomb xG: 0.9 - 1.2
Understat xG: 0.75 - 1.68
What Happened? A defensive lapse put United down inside the first 10 minutes only for United to level things up 13 minutes later. Pretty even from there on out with United getting the better of the chances, but not really any great ones (hence the difference between the statsbomb and understat xG values). A 65th minute Paul Pogba goalazo was the difference.
Luck rating: 4 - United struggled to break down a compact Fulham side and won when one of their special players did something special. Then again, that’s why you have good players, they tend to help you make your own luck.
West Brom 1-1 Manchester United
Statsbomb xG: 1.1 - 0.4
Understat xG: 0.95 - 0.42
What Happened? United were unlucky to concede a 2nd minute goal on what was a blatant foul against Victor Lindelof but if that one hadn’t counted West Brom would have found the net sooner or later. United were awful in this one and though they equalized before half time, the only team that looked close to scoring a winner was the soon to be relegated Baggies.
Luck rating: 6 - getting out of that match with a point is damn well impressive
Manchester United 3-1 Newcastle
Statsbomb xG: 1.7 - 0.5
Understat xG: 1.73 - 0.34
What Happened? The two teams emerged for the second half with the score tied 1-1. 11 minutes into the second half Dan James put United ahead. 20 minutes later Bruno Fernandes added a penalty. Chasing the match, Newcastle took just four shots in the second half, one was from a set piece and none were from inside the box.
Luck rating: Oh wait, I snuck this one in on you. United scored first in this one via Marcus Rashford but conceded an equalizer six minutes later. They went to halftime 1-1. Does that make the win more deserved because this time United scored the first of two first half goals?
Manchester United 2-1 Brighton
Statsbomb xG: 1.8 - 1.1
Understat xG: 1.73 - 1.41
What Happened? Horrible defending allows Danny Welbeck to open the scoring 12 minutes into the match when he hit in his own rebound. Welbeck’s initial shot and subsequent goal accounted for 0.95 of Brighton’s 1.41 xG. United’s attack was disjointed and a mess for most of this one, but Brighton offering absolutely zero threat meant by the half hour mark it was merely a case of ‘just get two goals over the next 60 minutes.’ The second half United sorta figured themselves out. They were still disjointed in attack but had complete control. The second half xG was 1.59-0.00. United created much better chances, missed a couple big ones, but also scored twice to take home all three points.
Luck rating: 3 - just awful defending on the opener but the chance only happens when Mike Dean completely blocks off Fred from the ball giving possession over to Brighton. Attack was a mess the entire game but again play this match over and over again and United come out on top a lot more often than not.
Tottenham 1-3 Manchester United
Statsbomb xG: 1.0 - 1.9
Understat xG: 0.92 - 2.10
What Happened? An Edinson Cavani goal disallowed by a controversial VAR decision meant United went 1-0 down when Spurs scored from their only chance of the match. United equalized 12 minutes after the restart and took the lead with 11 minutes left, before adding a third for good measure. It was one of their best performances of the season.
Luck rating: 2 - United were more unlucky than anything to even have gone down in the first place. One of their best performances of the season and one of those ‘you could sense the winner was coming just a matter of when not if’ feelings for 10 minutes prior to Cavani’s winner.
Aston Villa 1-3 Manchester United
Statsbomb xG: 0.5 - 1.8
Understat xG: 0.57 - 1.89
What Happened? 23 minutes in Bertrand Traore hits a nice strike, after McFred give away possession cheaply, not so much against the run of play, but certainly out of nowhere. United were in pretty good control of the match up to that point and maintained it the rest of the way. They leveled things up seven minutes into the second half when Paul Pogba won a penalty and took the lead four minutes later. They added another three minutes from time for good measure.
Luck rating: 1 - were playing well the whole way and were ahead after an hour
11 matches. That’s it. That’s the entire list. If you average out the very unofficial luck ratings it comes out to 3.99. Now I don’t know if that’s good or bad since it’s a very subjective number and I didn’t calculate this for any other team, but it doesn’t seem that bad to me.
Just because they’re coming back doesn’t make them exceedingly lucky. Cavani’s 91st minute winner at Southampton wasn’t even United’s latest winner of the season. That would be Rashford’s 93rd minute winner against Wolves - a match where they never fell behind. Mason Greenwood’s winner against Brighton came in the 83rd minute, two weeks later he scored an 84th minute against Burnley. In that match United scored first and then conceded an equalizer but in the 65th minute of both matches the score was 1-1. Does that make one more lucky than the other?
For all the late goals they scored they also conceded some late ones, throwing away wins against Leicester and Everton in the last five minutes. In both of those matches United scored first.
United were very consistent in how they played last season. In a season where everyone was exhausted, United let their opponents burn out while trying to take the game to them then ran them off the pitch in the second half7. This season it’s United who are faltering late all too often.
United certainly did get lucky a few times last season but luck has a funny way of evening itself out over the course of a season. But ultimately this team was just good, finishing right around where they should have finished.
There are several things wrong with Manchester United this season but one thing that didn’t happen was “they ran out of luck.” If anything I’d say they’ve been more lucky this season8 where they’ve certainly developed a habit of pulling results out of their ass.9
United did have a manager who was famous for consistently getting injury time winners and I don’t recall anyone ever calling it lucky or unsustainable
Which say what you will about it
Leicester’s title winners came in 12 points above their xPTS, City’s 2017-18 title winners were +8.91, while Liverpool’s 2019-20 title winners finished with 24 more points than their xPTS total - which was second in the league behind City. United’s runner up 2017-18 season saw them beat their sixth best in the league xPTS total by 18.67! On the other side was last season’s Brighton who played great football but lacked potent goal scorers up front, they came away with 20.41 fewer points than expected points had them winning.
This year they’re currently ninth according to both models so uhh yea, things have gotten worse
Oh shit, did I just turn that entire post into nothing more than a 1000 word introduction? You betcha!
One of which a James Ward Prowse worldie
Back in April I wrote about how this was happening so often it may have been tactical. United were playing within themselves in the first half before blowing teams away in the second half, like a boxer absorbing punches early in the fight and letting his opponent wear himself out before knocking him out later.
I purposely didn’t even touch on the Champions League because six games is a tiny sample that lends itself to tons of variance but:
Last season United drew a tough group yet still finished with one of the best NPxG differentials and didn’t go through. This year they drew an objectively easier group, had a far worse NPxG differential and went through. Their NPxG Diff is 12th best among the 16 teams that qualified. For the round of 16 they’ve managed to draw one of the four teams worse than them… decent luck there
Wolves away, Villarreal home, Villa away, Villa in the FA Cup, and even Crystal Palace at home ranks pretty highly on the ‘fortunate to get all three points’ scale