In All Probability

In All Probability

If you read my previous article, you will know that the pursuit of data driven betting excellence has once again got underway for the weekend’s Premier League action and although match outcomes are always an interesting thing to look at, there’s another aspect we should not leave behind.

GOAL PROBABILITY

One of the more widely tackled matters in terms of data driven football predictions is the Under/Over 2.5 Goals Market and that is the topic that we are now going to revisit for the rest of the season, as the data set has been replenished accordingly.

For those who tuned into this test last season, over 12 weeks of the season (pre-Covid) 82 Premier League fixtures hit a certain criteria to be bet on – as unlike the match outcomes, I do not test this on everything.

Of those 82, there was a 69.5% success rate (57 out of 82) and although the sample size was a little on the small side, it did at least give us something to work with this time around and with the mindset being if isnt’ broken, don’t fix it, we are going to try the same this season.

THE CRITERIA

For those who did not follow last season, there is no ‘magic’ to any of this, it is nothing more than probability – outcomes of the past, deciding what happens in the future and although that is far from a perfect science, the hit rate from last season suggests its not the worst logic idea.

While after waiting for the first seven weeks of this season to take place, it means that there’s enough data for us to be working with in terms of probability and in the first week of operation (week 8), here is what the data looked like:

Anything over 55% = Green (Bet on Over 2.5)
Anything between 45% and 54.99% = Orange (Ignore)
Anything under 44.99% = Red (Bet on Under 2.5)

So in the first raft of six bets, three out of the six were correct

THE FOLLOWING WEEK

With the same logic and criteria now in play, here is what the data suggested last week:

Last week, seven matches cleared the criteria and five were correct.

Which means for the season so far 8 out of 13 have got over the line = 61.5% and although this is an ever smaller data sample than in 2019/20, there’s plenty of opportunity for this to improve as the weeks continue.

WHAT ABOUT THIS WEEK

This week, eight games have cleared the criteria and these are the suggested bets.

One thing that is interesting is that Man City vs Burnley is such a low probability, but at the same time this is perhaps reflective of both team’s scoring problems thus far and the old occasions where Pep’s men would explode with goals are not quite the regular recurrence that they once were.

So with the data dumped, that’s all that needs to be said for this week. Do feel free to use these as a steer for your weekend buts and if you do use them, best of luck.

Happy punting and thanks for reading.

(THESE ARE NOT TIPS PER SE, THESE ARE JUST DATA DRIVEN FOOTBALL PREDICTIONS IN A PURE TEST AND LEARN MODELLING CAPACITY)