Rotate to accumulate: How the argument has turned full circle
Daniel Finkelstein, The Times
It's a constantly fascinating thing this Fink Tank business. The gift that keeps on giving. I am learning all the time. What have I learnt this week? Something about rotating players.
A few years back, Claudio Ranieri, then Chelsea head coach, acquired the nickname “The Tinkerman” because of his inability to pick 11 players and stick with them. We set about studying whether this helped or hindered his team.
How did we measure this? By looking at the percentage of the time his, and other teams', top 11 players were on the pitch on average. The lower the average percentage for each side, the less consistent their team selection.
We then tried to see whether there was a relationship between this number and the number of points the teams gained. We took account of the money each team had and then did our calculation.
Well, well, well. It turned out that inconsistency was a bad idea. And we've just repeated that work for the last few seasons and demonstrated that it remains a bad idea. How bad? It was about a third as important as the wage bill. In other words - pretty important.
Then, a couple of weeks ago, we decided to return to the subject. Rafael Benítez is being accused of rotating his players too much. Just using our old Ranieri analysis, one might conclude that this was a bad thing. But was using this old analysis and applying it to Benítez the right thing to do? Dr Henry Stott and Dr Ian Graham decided it was not.
The accusation against Benítez is not that Liverpool don't use their top XI enough over a season, it is that he darts around from game to game, resting players for a match or two.
You can measure that behaviour by looking at how many players on average are the same in one game from the game before. And when you do, you can see that consistency and rotation are related but not the same thing. Last season, for instance, Aston Villa rotated very little but had an averagely consistent team selection, while Chelsea were the opposite. Liverpool? Averagely consistent, rotated a lot.
What happens when you look at rotation and its relationship with performance? The first step is to find out which clubs rotate a lot and which rotate little. The graphic shows last season's measure. The fewer the number of players left unchanged from match to match on average, the greater the amount of rotation.
And as can be seen, Liverpool are at the top end with Manchester United and Chelsea, but are not the biggest rotators. Indeed, any statistician used to dealing with means will observe that the differences between them are within the margin of error - essentially that they are the same.
Next we related the rotation measure to points gained, again taking the wage bill into account. And what do we find? That rotation is related to gaining more points, not fewer. How strong is this relationship? Well, a few years back pretty strong. More recently the wage bill has become such a strong determinant of success that the impact of rotation, while still positive, has become less important.
You need a balance. You don't want to be inconsistent (as Ranieri was), you need a top 11 playing a good deal of the time, but you do want to rotate (as Rafa does).
Rotation has been identified as a culprit when teams lose, but it is not to blame. United rotated more than Liverpool last term and won the title.
Link
Daniel Finkelstein, The Times
It's a constantly fascinating thing this Fink Tank business. The gift that keeps on giving. I am learning all the time. What have I learnt this week? Something about rotating players.
A few years back, Claudio Ranieri, then Chelsea head coach, acquired the nickname “The Tinkerman” because of his inability to pick 11 players and stick with them. We set about studying whether this helped or hindered his team.
How did we measure this? By looking at the percentage of the time his, and other teams', top 11 players were on the pitch on average. The lower the average percentage for each side, the less consistent their team selection.
We then tried to see whether there was a relationship between this number and the number of points the teams gained. We took account of the money each team had and then did our calculation.
Well, well, well. It turned out that inconsistency was a bad idea. And we've just repeated that work for the last few seasons and demonstrated that it remains a bad idea. How bad? It was about a third as important as the wage bill. In other words - pretty important.
Then, a couple of weeks ago, we decided to return to the subject. Rafael Benítez is being accused of rotating his players too much. Just using our old Ranieri analysis, one might conclude that this was a bad thing. But was using this old analysis and applying it to Benítez the right thing to do? Dr Henry Stott and Dr Ian Graham decided it was not.
The accusation against Benítez is not that Liverpool don't use their top XI enough over a season, it is that he darts around from game to game, resting players for a match or two.
You can measure that behaviour by looking at how many players on average are the same in one game from the game before. And when you do, you can see that consistency and rotation are related but not the same thing. Last season, for instance, Aston Villa rotated very little but had an averagely consistent team selection, while Chelsea were the opposite. Liverpool? Averagely consistent, rotated a lot.
What happens when you look at rotation and its relationship with performance? The first step is to find out which clubs rotate a lot and which rotate little. The graphic shows last season's measure. The fewer the number of players left unchanged from match to match on average, the greater the amount of rotation.
And as can be seen, Liverpool are at the top end with Manchester United and Chelsea, but are not the biggest rotators. Indeed, any statistician used to dealing with means will observe that the differences between them are within the margin of error - essentially that they are the same.
Next we related the rotation measure to points gained, again taking the wage bill into account. And what do we find? That rotation is related to gaining more points, not fewer. How strong is this relationship? Well, a few years back pretty strong. More recently the wage bill has become such a strong determinant of success that the impact of rotation, while still positive, has become less important.
You need a balance. You don't want to be inconsistent (as Ranieri was), you need a top 11 playing a good deal of the time, but you do want to rotate (as Rafa does).
Rotation has been identified as a culprit when teams lose, but it is not to blame. United rotated more than Liverpool last term and won the title.
Link
Comment