With my last article I introduced the concept of Goal Rating Models. I showed you that a very simplistic rating model won’t guarantee you profit. However, tweeking and adjusting your model will make it better in time and get you closer to where you want to be. In order to do that, you need to understand the stats your model is based upon – in this case, goal stats. To help you with this I wanted to present you one of the anomalies typical for the football market – the rapid increase of goals scored in the last rounds of the season in most European Leagues. Incorporating such anomalies into your model will surely improve its predictive power.
Goal Stats: End of Season Frenzy
Today I share with you one of the interesting observations I have made when researching European goal data from the last ~20 seasons regarding goal statistics. As some of you might know, the average number of scored goals in the last rounds of the season increases dramatically for just about every major European football league (which is probably true for all leagues in general, although I don’t have data to confirm it). Have a look at the average number of goals per country for the last 15 rounds of the season, calculated over the last 22 seasons using the available data at football-data.co.uk.
You see there is a clear upward slope towards the end of the season. The increase is less pronounced for England and Scotland and more for the continental leagues. For Italy, Germany, France, Spain, the increase seems to start 9-10 rounds before the last one, to peak in the very last round of the season when the most goals are being scored. There might be different explanations for this phenomenon (tired players, teams lacking motivation resulting in a less tactically sound game), but it’s certainly there.
I got pretty excited when I found out about this anomaly. As it is known that on average more goals are scored in the second half of a football game than in the first, it seems to be another universal rule that more goals are scored in the last weeks of a season than in the rest of it. Of course after finding out about it the first thing on my mind was to see if a profitable strategy might be built around it.
Do the markets know about it?
I decided to omit England and Scotland from the test since the increase in goal scoring at the end of the season for those leagues was less profound then for the others. Moreover, as I noticed the increase starts around 10 rounds before the end of the season, I only tested a strategy backing the >2.5 goals at the maximum available odds for the respective league in the last 10 rounds of the season. I have calculated the averages from the last 11 seasons (as this is the current span of odds data at football-data.co.uk). Here are the results:
The first letter stands for the country (Spain, Italy, Germany and France) and the number for the tier of the league, first or second. One can clearly see that from the 8 tested leagues the strategy would have been profitable only for Segunda, Ligue 1 and somewhat for Serie A. Moreover, it is striking how low the returns for the German leagues go in the last round(s) considering that Germany is the country with the most profound increase in goals and the most goals in general as could be seen in the first graph. On the other hand, although France appears to be the league with the least goals among the tested, the profit backing overs there seems to be one of the higher ones.
Can you make a profit with it?
Despite the fact that for Segunda and Ligue 1 one can see a nice steady increase in profit, the strategy for all leagues combined seems to be a losing one. Whether it is worthy to try it for those two leagues alone you can decide for yourself. In any case, I think it is a good example for how a trend in the stats is already incorporated in the bookmaker’s price before we even know about it. Therefore stats alone don’t tell us anything in the betting world unless they are viewed from the perspective of the implied probabilities as represented by the betting odds.
Goal Stats: What are they good for?
Obviously, one cannot make a profit by betting every trend observed on a market. However, if you model a rating throughout the season, you should know about those anomalies. Otherwise you will weigh a goal at the end of the season the same as one at the beginning. However, obviously, those are not worth the same. A goal in the beginning of the season might secure you important points in the midst of a battle for the title or for avoiding relegation. A goal at the end of the season likely comes from a meaningless game, where everything has already be settled. Those are mostly only slightly more important than a regular friendly game. This, in turn, is reflected by the number of goals scored in them. Obviously, not all goals are made alike and the Goal Stats show that.
If you want to use this information to build a Goal Rating Model, I strongly suggest to refer to my step-by-step guide on Building a Rating Model, as well as check my last article for a practical example.
I am planning to publish a few more articles on interesting football stats so if this is what you are looking for make sure to follow me on Twitter or Facebook in order to be notified for new blog posts. Thanks for reading and see you around!