Euro Basketball analysis: h2h

Model, Sport Models | | January 8, 2015 at 3:13 pm

hoFznCKE323ngrlERnY8-mSportpunter’s Euro Basketball model has been in operation for free since 2011, and we thought it’s about time that we gave you some analysis of how it has gone. This is the first of a three part article on the statistical analysis of the results. This one will be about head to head betting, whilst the next will be line and then totals betting.

So let’s get started. Shown on this link is a summary of all the data for European Basketball, betting h2h. It shows that the model has made a conservative 2% ROI from over 6000 bets. But when we look a little closer, we notice that it has not done well on underdogs at all. In fact, it has lost 4.9% ROI betting on teams where were rated them less than a 40% chance to win. Conversely, it has made 4.7% ROI from 3400 bets where we rated the team a 40% chance or greater.

Note also the big losses when the odds were over 6. Using a minimum overlay of 12.5% seems to be optimal, and it would seem that there is no minimal overlay to use on underdogs. So suggested bets on teams with probabilities less than 40% would be best to be ignored.

A 6.8% ROI would have been made since 2011 had you bet on European Basketball teams with probabilities greater than 40% with a minimum 12.5% overlay.

Of course, this is very selective, and it’s only because we know the past data that we can analyse it in this way. However it is useful for making a scheme to bet in the future. One has to have a logical reason why the underdogs don’t seem to be winning. If we were to bet blindly on favourites to win $100, then we would have lost 2.2% ROI. However if we were to bet blindly on underdogs, we would have lost 2.1%, so this is quite comparable. However, if we bet to win $100 on every team with odds of greater than 6, we would have lost 17.2% ROI from 842 bets.

So clearly the favourite/longshot bias is there in the sport, and it would take a very good model to beat the 17% advantage that the bookmakers have at these high odds. I also have a theory that the model is bound to be conservative for big outsiders because of the assumptions of their distribution.

Shown here is how the model has gone betting on each league, which shows Croatia losing significantly, whilst France LNB Pro A, Spain ACB, and WNBA have had great returns. It might be wise not to read anything too major into these results due to small sample sizes.

Either way, whilst the overall percent return is a modest 2%, one can make significantly more by not betting on large underdogs as well as increasing the minimum overlay.

Our free European Basketball predictions, betting history and analysis can be found on this link.

Related Posts with Thumbnails

Leave a Reply