Sack the manager! Or should we...

As of kick off on Saturday, the bottom 7 Premier League clubs had all changed their manager this season, while Spurs were the only club of the top 13 to do so. At first sight, that sounds like a new manager is a terrible idea. And indeed, many would agree. After all, has Ole Gunnar Solskjaer really done more with Cardiff than Malky Mackay? Fulham are already on to their third manager and might not be the only club to sack the first replacement.

On the other hand, Tony Pulis has been, rightly in my opinion, credited for keeping Palace up while Pepe Mel is increasingly getting results at the Hawthorns. Obviously the Norwich hierachy subscribe to this view, deciding to sack Chris Hughton prior to Saturday's game at Fulham. It hasn't worked so far...

So, does a new manager mean a profitable honeymoon or certain doom?



Let's look at the immediate impact first. The table below shows the Premier League managerial changes so far this season, with the last result under the old manager and the first result under the new manager (note this might not be the next result, as there can be a gap between appointments).

As we can see, for most clubs the supposed "new manager" effect doesn't seem to really take hold. Of the 9 managerial changes thus far, only 2 have resulted in a win in the next game. But there is a problem with such a direct comparison - teams don't necessarily improve overnight. And furthermore, this doesn't really take account of the vagaries of the fixture list. After all, is it fair to compare a single game with another, when one could be Man City away and the other Sunderland at home?

Let's zoom out a bit, and have a look at the five games either side of the managerial change. I'm including Cup results and assigning the same points as in the League, for ease of comparison. I've left Norwich out as we can't analyse Neil Adams first five games, given he's only been in charge for one!

When laid out like this, we can see that, with the arguable exceptions of AVB and Laudrup, most teams are in unsurprisingly awful form before getting rid of their manager. We can also see that most seem to improve. Let's plot that on a graph, to make things clearer still.

So it looks like a considerable improvement in the immediate 5 games following the change. In fact, on average teams get an extra 3 points from the 5 games under a new manager compared to the final five for the old one. So far, it looks like Norwich have made the right decision.

But there's still two problems with our approach so far. Firstly, while we appear to have shown there is a "new manager" effect in the first 5 games, we have not shown whether this effect lasts. So we need to consider managerial performance across the rest of the season. Secondly, as the table shows and we know anyway, managers tend to get sacked when the team is doing badly. Even five games can be an unrepresentative blip in a season of at least 40 games in all competitions. So we are now going on to compare the points per game (PPG) achieved from the start of the season until the outgoing manager leaves, with the PPG achieved by the new man up to today. We're also going to expand our data by looking at the last two seasons as well as this year.

So it looks a lot like changing manager during the season is worthwhile - on average, teams achieve 16% more points per game after the new man is appointed. If we apply the average improvement in points per game, this might save Norwich from the drop - they could expect to get at least 36 if not 37 points, rather than the 35 maybe 36 they were due to get beforehand. That could be the difference in staying up...although it doesn't account for their difficult run-in.