## Are goals scored in the first half of a game predictive of goals scored in the second?

Somehow I haven’t posted this before but I was wondering whether a high scoring first half of a game is predictive of goals scored in the second half so I’m going to look at that in a few different ways.

Firstly, here’s a breakdown of the half time scores in the last 5,100 Premiership games – both in terms of raw numbers and percentages.

These games average 1.16 goals in the first half. The average game in this sample has 2.64 goals, leaving 1.48 to be scored in the average second half. However, does the number of goals scored in the first half have any bearing on the number of goals scored in the second?

Firstly I’m going to split games by the score at half time (home team leading/game tied/away team leading):

Home teams score 57.8% of Premiership goals but it turns out the split isn’t exactly even between the first and second halves – they score 58.5% of first half goals and 57.2% of second half goals. I’d hazard a guess at score effects but I’m definitely open to other explanations here.

There’s two further very noticeable splits here – I think they’re pretty explainable though. The first is between tied games that average 1.43 second half goals per game, and not tied games that average 1.51. I think this is because there’s less incentive for tied teams to go after a goal than losing ones – they’re both guaranteed a point if nothing happens the rest of the game. I imagine if I had the breakdown of shots in each half of the game available to me that we’d see a very similar pattern.

Secondly the percentage of second half goals scored by the home team drop from 59.6% in games the home team leads at half time, to 57.2% in games that are tied at half time, to 53.7% in games the away team leads at half time. This one seems pretty obvious – the better team is more likely to be leading at half time and score a greater proportion of second half goals simply as a result of being the better team.

Next I want to look at the number of goals scored in the first half, and see how they impact upon the number of goals scored in the second half. I’m only including games with 0-4 goals first half goals, as above that the sample sizes drop below 38 games – my arbitrary cut off point, but below that level we’re considering less than 0.8% of the sample.

So there’s a moderate increase there – we expect to see ~25% more goals in the second half of a game that features 4 goals in the first half than in a game that’s scoreless in the first half.

So lets go one more level, break down the half time results by score, and see the average number of goals scored in the second halves of those games. Below is a table of those half time scores – and the average number of second half goals in those games. Once more I’ve only looked at scorelines that have a sample size of at least a seasons worth (38) games.

And to me that’s a pretty nice illustration of the increase in second half goals scored as both the number of first half goals increase, and the further a game is from being a tied score.