Man United’s frustrating Old Trafford draw with West Ham last Sunday week was their fourth succesive home draw, and a pattern that has not been replicated since 1980. Coincidentally it was also the year the American public elected a pseudo-celebrity to become President of the United States. With Hammers goalkeeper Randolph making a series of stunning saves, and similar happenings against Burnley and Stoke, post-match, Ander Herrera was convinced the Red Devils had done enough to turn those draws into wins. He remonstrated, “I don’t know what to say. At the moment, luck is completely against us. We are playing very good and we deserve eight more points than we have.” He continued, “I don’t know when the luck will change, but it’s not even luck - it’s just (what is) fair. We are looking for fair. The best player of Burnley was the keeper, the best player of Stoke was the keeper, the best player of Arsenal was Petr Cech, the best player of West Ham was Randolph.” And finally, “It’s unbelievable that we don’t have six or eight points more than we have but hopefully the luck will turn in our favour.”

It poses an intriguing question. Have Man United been unlucky this season? While separating luck from skill is difficult we can make some light assumptions based on certain longer term trends, and more specifically by looking deeply at chance quality. As the Spaniard’s quotes mainly refer to man of the match accolades going to goalkeepers that Man United have faced, it would be more prudent to narrow down the chance quality to just shots on target, and as a consequence using *XG2 (chance quality with shot placement taken into account) will add more context when dealing with goalkeepers.

Extracting shots on target from the Burnley, Stoke and West Ham games I ran a Monte Carlo simulation (20k iterations) on Man United’s shot placement chance quality values (XG2). I’ve left Arsenal’s game out of the equation as I’m not sure I agree with Herrara in that they SHOULD have won that game. Though they had the better of the play and better all round possession they didn’t create a single OPTA defined big chance against Arsenal and only slightly bested them in terms of chance quality. (Disclaimer: single game evaluations of chance quality may be open to variablility. End Disclaimer.) Note: disclaimer works best when read with robotic voice.

Here’s the resulting histogram:

Simulation


In only 65 of those 20k simulations (0.003%) an average team would score just 2 goals. In fact the average team given those chances with that shot placement are more likely to score 12 goals than 2. It’s also worth noting that Man United’s average chance probability (XG2 per shot) for each shot on target for those specific games was ~30%, where the average for a team in the Premier League is closer to ~22% per shot. Of the 112 games last year where a team was expected to have scored 2+ goals based on combining their chance quality and shot placement there were 19 of those teams that produced just 1 goal or less. It happened to no individual team more than twice. It’s already happened to Manchester United 3 times this season.

Visually, here’s a look at Man United’s shots on target against Burnley, Stoke and West Ham this season.

Visual

This is pretty high chance quality shot placement, with the 2 shots that were scored highlighted. 11 of these shots had greater than a 35% chance of becoming a goal, and 4 of those greater than a 66% chance. Whereas on the other side of the coin it’s reasonable to say that each goalkeeper did an extremely good job at keeping out those shots.

Players

Moving on, in an overall context how are Man United’s players doing in terms finishing? Here are the players who lead the way in cumulative chance quality for Man United this season.


Zlatan, Rooney, Pogba and (slightly) Rashford are all under-performing in converting their chances (**goal efficiency) this season. Though Rooney’s history suggests that finding the net hasn’t been a problem for him in the past. Since 10/11 where he slightly under-performed his chance quality, he has over-performed 5 years in a row. Ibrahimovic on the other hand has been a touch more patchy, under-performing in 2 out of the last 4 seasons. Whilst Pogba has managed over-performing in 3 out of the last 4 seasons. Finishing. Is. Hard.

Using a quicker way (rather than simulations) of ranking these players finishing I used Peter Owen’s Binominal Ranking method (for a number of reasons one of my favourite pieces) to assess these players. (minus Marcus Rashford due to his age and sample size issues) This is using all data I have on file for each player, which contains 500+ shots for both Rooney and Zlatan, and 250+ shots for Mata and Pogba.


Besides having probably the best shot selection I’ve come across over the years, Rooney comes out as immense over his career. Mata, Zlatan and Pogba have all had fairly decent finishing numbers throughout these seasons. Whilst any one player can have a bad season in terms of finishing it would seem given their history these players are more likely to over-perform their chance quality than under-perform. It should be said, we are still within a relativealy small sample period, and with just 13 games gone of the new season it may be the case that this is just random chance rearing it’s head. In other words, there is plenty of time for these players to ‘catch up’ on their chance quality. Or get worse!

In concluding Man United have been somewhat unlucky this season. Which isn’t any great revelation, but it’s nice to look at the numbers behind it. Whether, as Herrara has suggested, they’ve been 6-8 points unlucky I’m not so sure. So to put a number on it I ran another Monte Carlo simulation based on my chance quality model on all the games played so far this season and it indicated a 5 point deficit (with a SD of +/- 1.9 points) to what they should have. Which wouldn’t change their position in the table, but would put them within a point of Spurs and 4 points off those coveted Champions League places. It’s been fun, thanks for reading.



————————————————————————————————————————————————————————————————–

*XG2: as well as looking at all the variables (shot location, shot type, big chance, shot after a dribble, counter attack shots etc etc) that go into chance quality (XG(1)) the added factor of shot placement in the goal is taken into account. For example a shot in the box from wide left of goal, aimed at the far right top corner will have a higher likelihood of becoming a goal than one from the same position aimed at the near post. (Closer to where the goalkeeper might position himself)

**Goal efficiency is Goals/expected goals. Anything below 1.0 is under-performing, whilst over 1.0 is over-performing. 1.0 would be scoring at the rate an average player would given the chances they had.