Sunday, May 1, 2016

Pellegrini's Lack Of Rotation Is Hurting Manchester City

In any discussion of Manchester City’s season, one of the topics that inevitably come up is injuries. There is no doubt that missing Kompany, Aguero, De Bruyne, Silva, and Nasri for various parts of the season has had a negative impact on the team’s performance. Conversely, one of the primary reasons for Leicester’s success has been the lack of injuries to their starting eleven. Unfortunately, most pundits and fans don’t go further than lamenting the injuries, leading to articles like “De Bruyne’s injury cost City the title” (a statement that conveniently ignores that Manchester City’s best performances came in August without De Bruyne). However, that implicitly assumes that injuries are simply an external factor in the hands of the gods and not something teams have control over. While luck is certainly involved in injuries, there is increasing evidence (both in football and other sports) that providing players enough rest can reduce the risk of injury. In this regard, Pellegrini’s rotations throughout the season have not helped matters and he deserves culpability for his role in City’s injury crisis.

The science of what causes specific injuries still remains the Holy Grail for analytics researchers across sports. Knowing what types of players are going to experience more injuries can prevent clubs from overpaying the likes of Andy Carroll or Daniel Sturridge (or even, and I hate myself for writing this, Vincent Kompany), but they are still hard to identify prior to their injury troubles. However, from an aggregate basis, it has become obvious that providing enough rest to players is critical in injury prevention. This study (described here) of a team participating in the UEFA Champions League in addition to their respective league shows that when players have four or fewer days of rest, they are 6 times as likely to have an injury as when they have a full week. One of the reasons it is not surprising Leicester has done so well on the injury front is they have no European games and didn’t have deep runs in either domestic cup. As such, they rarely had to play midweek games and were able to provide more weekly rest as a result, which the study suggests would dramatically reduce injury risk. Studies in other sports confirm the same result: this report shows changing attitudes towards rest in the NBA and how teams are rotating their players in an effort to avoid injury. The message is clear: the more rest a player has between games, the less likely he is to get hurt.

Unfortunately, Pellegrini doesn’t seem to have gotten this message. By going through lineup data and injury data from* for each game (up until 4/12/16), I have created the below table. It shows the number of enforced and unenforced changes to Manchester City's lineup for each game, the days of rest since the last game (I put in 6 for the first game of the season), and the total number of those injured or suspended for the game. If Pellegrini was rotating players properly, we'd expect to see more unenforced changes when the team has had fewer days of rest.

Date Comp Enforced Changes Unenforced Changes Days Rest # of injuries
8/10/2015 PL 0 0 6 4
8/16/2015 PL 0 1 5 3
8/23/2015 PL 0 0 6 1
8/29/2015 PL 0 0 5 4
9/12/2015 PL 1 1 13 5
9/15/2015 UCL 1 1 2 4
9/19/2015 PL 2 2 3 5
9/22/2015 CC 1 2 2 7
9/26/2015 PL 0 1 3 7
9/30/2015 UCL 0 2 3 5
10/3/2015 PL 1 2 2 5
10/17/2015 PL 3 1 13 6
10/21/2015 UCL 0 0 3 5
10/25/2015 PL 0 3 3 5
10/28/2015 CC 1 5 2 6
10/31/2015 PL 1 4 2 6
11/3/2015 UCL 1 1 2 5
11/8/2015 PL 0 1 4 4
11/21/2015 PL 2 2 12 6
11/25/2015 UCL 1 2 3 5
11/28/2015 PL 1 3 2 5
12/1/2015 CC 1 4 2 6
12/5/2015 PL 1 4 3 6
12/8/2015 UCL 2 2 2 6
12/12/2015 PL 0 3 3 5
12/21/2015 PL 0 4 8 4
12/26/2015 PL 0 2 4 4
12/29/2015 PL 0 1 2 4
1/2/2016 PL 0 0 3 2
1/6/2016 CC 0 3 3 3
1/9/2016 FAC 0 7 2 3
1/13/2016 PL 0 5 3 5
1/16/2016 PL 0 5 2 6
1/23/2016 PL 1 3 6 6
1/27/2016 CC 0 4 3 5
1/30/2016 FAC 0 4 2 6
2/2/2016 PL 1 4 2 6
2/6/2016 PL 1 3 3 6
2/14/2016 PL 1 2 7 7
2/21/2016 FAC 0 9 6 7
2/24/2016 UCL 0 10 2 6
2/28/2016 CC 0 1 3 3
3/2/2016 PL 1 2 2 4
3/5/2016 PL 0 3 2 3
3/12/2016 PL 1 0 6 4
3/15/2016 UCL 0 2 2 3
3/20/2016 PL 2 2 4 5
4/2/2016 PL 3 2 12 5
4/6/2016 UCL 0 2 3 3
4/9/2016 PL 1 4 2 3
4/12/2016 UCL 0 5 2 2
Averages 0.63 2.67 3.96 4.73

Not noticing that pattern? That's because it doesn't exist: there is no correlation between fewer days of rest and more unenforced changes (the r-squared is .05). I thought the international break might be skewing things, for although that leads to two weeks between games for Manchester City, most City players have multiple games in between and Pellegrini might be rotating more on those days as a result (the impact of international games on club sides is one reason I hate international football). However, when I removed matches with greater than 10 days rest from the sample, the correlation actually gets worse (r-squared of .02). I thought it might be that Pellegrini rotates a lot less when there are more total injuries, but there's actually a slight positive correlation between the number of unavailable players and the number of unenforced changes. Put simply, Pellegrini is not rotating his squad adequately to give his players enough rest.

Are other teams doing better at managing player rest? The short answer is "Yes". The longer answer is "Yes, but at least we're doing better than United." I looked at the same data for Spurs, United, and Arsenal as well, all teams with European commitments in a similar position in the table. A comparison between the four clubs is below:

Team Average Enforced Changes Average Unenforced Changes Average Days Rest Average # of injuries R-Squared Rest/Rotation
City 0.63 2.67 3.96 4.73 0.05
Spurs 0.60 3.60 4.27 3.17 0.25
United 0.56 1.98 4.06 6.16 0.04
Arsenal 0.34 3.04 4.34 5.77 0.14

Spurs are quite clearly leading the way here, though Arsenal look better as well, with a much stronger correlation between the number of days rest and the number of unenforced changes. One could argue that since they are the only one of the clubs playing in the Europa League, they are able to rotate more easily for those games, and they do have an average of 3.7 unenforced changes for those games, compared to 2.7 for City's Champions League and 1.9 for United's. However, Arsenal had an average of 3.5 for their Champions League games, and Spurs progress in the tournament shows they were not taking it lightly. What's really stunning about their strong rotation policy is that they are one of the teams that least need it, as they have one of the youngest squads in the league. The youth of the squad and good rotations are likely helping them keep the number of injured below the other teams on the list.

The most interesting comparison though is with our neighbors in Trafford, as the teams' approaches are very similar. As with Pellegrini, Van Gaal has displayed no concern for resting players when they have had fewer days rest and has even averaged fewer unenforced changes per game this season. However, there is a major difference: the large number of injuries has meant Van Gaal has plucked players from the academy to play for him. The amazing thing is that he doesn't even rotate those players, preferring to just call up the next one when the first goes down, a measure of trust that seems to belie the fact that none of the players were originally deemed good enough to be in the squad . This can clearly be seen in United's squad lists, with well-defined eras of previously unheard-of players such as the Borthwick-Jackson Epoch (where he started five straight games), the Varela Period (where he started 7 of 8) and the recent Rashford Renaissance. Pellegrini had similar options to promote youth, but except for Iheanacho (or, depending on your point of view, including him) he has limited young players to token appearances.

Are there any valid reasons that Pellegrini has limited squad rotation? Obviously, the importance City's hierarchy has placed on results in Europe is part of it, Pellegrini is afraid to do too much rotation in the Champions League. There is also an argument to be made that continuity can be important to helping a squad perform at its best. As a player, you develop knowledge of how to play with your teammates by increased match time together, a set squad allows for that trust and familiarity to develop. It's an argument that Pellegrini seems to buy, most evidenced by his preference to even have his backup goalkeeper play multiple games in a row when he is given the chance to start. That desire for continuity though is in conflict with the simple fact that a club in Manchester City's position will often have multiple games per week. It's not as though Pellegrini doesn't understand that rest has value, he argued to move the Southampton game to Saturday to increase the rest time prior to the Madrid game and complained about fixture congestion prior to the FA cup game at Chelsea. However, he seems to save his energy fighting for something that will not change instead of doing what he can to control the situation. With an old squad and a large number of midweek games, City should be rotating more than anyone, but Pellegrini seems to view having a large squad as cover for injuries, rather than as a way to prevent them. Though obviously no injury can be proved to be a direct result of lack of rest, Pellegrini is endangering his players by not rotating enough, and it's hurting them and Manchester City.

*There are a few games missing from this database. For those games, I assumed players injured for the previous game were still injured unless they appeared in that week's starting lineup.