By Jared Young (@jaredeyoung)
If you like goal scoring then tuning in for the 2nd half of MLS games will bring you 32% more pleasure than watching the 1st half. There have been 478 goals scored in the 2nd half versus just 363 in the 1st half through September 29th. Finishing rates improve from 10.3% to 10.9% between halves but the primary driver of the increased goal scoring is a 24% increase in shots attempted. Why do shots increase so much?
There are three potential answers to that question that I come up with and they deal with the physical, the mental and the tactical. 1) The physical is perhaps the most obvious. It's possible that players are getting tired and not maintaining their position or able to make defensive plays as frequently. The other physical option is that the second halves of games run slightly longer than the first half. This should account for a few percent of the increase. 2) Mentally perhaps players become less conservative as the game wears on and become more liberal with their shot selection. 3) The third option is that teams make tactical changes that change the rate at which teams shoot. The biggest cause of a tactical change is a change to the game state. When one team takes a lead it leads one team to be more aggressive which opens up the game. But tactical changes that open up the game may not simply be a matter of game state and could also influence shot attempts. So is it the physical, mental or tactical?
Measuring how tired a player gets is not particularly easy, especially without detailed GPS tracking data, so let's see if we can answer this question by solving for the other options first. We can examine the mental hypothesis and see if teams settle for less good shots as games wear on by looking at shot location and the very quick answer is yes, but not in the direction we're looking for. Shots attempted in the 2nd half actually get 2 to 3 feet closer to the goal. Expected goals models back this up as expected goals per shot increase from 10.7% to 11.1% after the break. This would explain the improvement in finishing rate but it does not explain how more of those shots are attempted. It does not appear that more liberal shot selection is a cause of the goal scoring increase.
Tactically speaking we can examine the impact of game state. Here is a chart of minutes per shot attempt by various goal states.
When there is a goal differential in a game then shots do happen more frequently, more than 10% higher when there is a difference of two goals. However, none of these data points reflect a 24% change in shots attempted. And this is obvious but there is a greater chance of a non-even game state as the game progresses. See below.
So game state pretty clearly influences shots attempted by opening up the game (and potentially resulting in better shots attempted) but game state also pretty clearly is not the complete answer.
Given that more liberal shooting is not the answer, the remaining gap must be caused by two things - players getting tired or tactical changes not related to game state. If players getting tired is at all a factor then it would seem to be something that is very correctable on a team by team basis. In fact, looking at the teams shows a very wide range of success between the first and second half in terms of goals allowed.
Four teams actually perform better in the 2nd half while the rest all perform worse. Columbus, Colorado and LA have all given up twice as many goals in the second half compared with the first. Do those teams open the game up more in the 2nd half or are they tiring?
I'd be very curious if there is typical of other leagues around the world or if this second half excitement is specific to MLS.