How xG can beat the bookies
It has its knockers but here is the proof that Expected Goals (xG) can be used to beat the bookies
You cannot fault Jeff Stelling for his passion but like many of the old pros that currently carve out a living analysing football matches, his refusal to embrace revolutionary new stats that enable us to deepen our understanding of the game remains mind-boggling.
One of these stats ‘xG’ (Expected Goals) was on the receiving end of a Stelling rant last November when he described it as a useless stat and absolute nonsense.
Jeff Stelling shared his thoughts over the 'Expected Goals' statistic at the weekend 😡😂 pic.twitter.com/SP2MgMuuoQ
— Soccer AM (@SoccerAM) November 21, 2017
But Stelling is not the only one that has been dismissive about the value of xG.
Craig Burley, who had the indignity of being sent off while playing for Scotland at the 1998 World Cup finals, was almost seeing red again when the topic of Epected Goals roared its ugly head while he was discussing a Champions League round on ESPN.
Many daily fantasy football managers are likely to disagree though. Understanding the value of underlying stats is a vital component of the player selection process and a skill that needs to be acquired if you are to have any chance of long-term success.
So where are we going with this you may ask?
Well, since we appreciate how important xG and even xA (Expected Assists) are when it comes to analysing players for fantasy football selection purposes, we were curious to find out if these stats could actually be used to help us to predict the outcome of soccer matches.
Now we all know, that when it comes to beating the bookies, it is RESULTS that matter. Punters should always look for an ‘edge’ when assessing any betting tips and this is exactly what we were looking for.
Regular visitors to the this site will know that we produce projected goals forecasts for upcoming games based on the latest odds available with the bookmakers.
So, what we decided to do was to model xG figures for upcoming games that were calculated from historical results against our projected goals forecasts, in an attempt to highlight potential value bets.
If there was significant differences, we would then shortlist the selections in question before checking to see if other key ‘influencing factors’ were satisfied before actually pulling the trigger on a ‘pick’.
These influencing factors would include motivation and the actual confirmed line-ups as well as team strengths such as set piece ability, an area that we feel will become even more significant in the future and if you don’t believe us we strongly advise you to check in over at Statsbomb and discuss the matter with them!
We ran the trial from January 2018 to the end of the last domestic season focusing on Europe’s top leagues. However, to ensure that everything was above board we also needed a respectable website where we could ‘proof‘ these selections to gauge how successful or unsuccessful they were, so we opted for the brilliant FansUnite platform and posted our picks under the alias xGTips.
We targeted 2 specific markets – the Asian Handicap lines and the Total Goals lines, submitting a number of picks and the overall breakdown of our results is shown here.
A ROI of over 26% was an astonishing performance and every league ended up in profit. However, we have been around the block long enough to know that such margins are not sustainable long-term and successful sports bettors that target these types of markets would be more than happy with a ROI of anywhere from 5-10%.
The great thing about this platform is that it records the prices that were available when the picks are submitted so there is no manufacturing of results that is common across many tipping websites.
We were therefore delighted to finish 80.19 points ahead from a turnover of 307 points!
So, with the new season almost upon us, we encourage you to embrace xG as your friend and ignore what the sceptics say – it is here to stay!
But….if you are in the Stelling or Burley camp we would love to know why?