Step By Step

Step By Step

Step By Step

In my previous article,  I referenced the fact that my pursuit of data driven betting excellence had taken me to more sunnier climbs and although I will definitely be keeping an eye on Spain’s La Liga this weekend, it is now time to go back to my roots.

The English Premier League may have a split gameweek but that doesn’t mean there is any less action (it just means we have to wait longer for it to complete) and with that in mind it is time for another raft of data driven prediction.


Now over the past few weeks, I have either had some rotten luck or just been rotten (to be honest it has probably been the latter) However, those poor performances have finally led me to making some fundamental changes and they are changes that should see me in good stead this time.

Last weekend, saw just four correct picks out of ten from ‘PremBot’ and I referenced that the logic meant that the majority of teams were cancelling each other out. Something that would indicate all was not well.

Of course, with such a poor return by the end of last weekend it obviously was not and therefore I have changed how the predictions are generated and followed the lead from the Spanish model in the previous article.


The main issue was I was not giving enough weight to dominant teams who are away from home and the model in its previous guise was coming back with some rather indifferent output, something I’ve managed to change by the following steps.

  1. If the home team is more or equal to five places higher in the league table and is much better at home, than their opponents are away then it is a home win
  2. Which before would have to be the same for an away team (Although much better on the road, then their opponents would be at home)
  3. However, fundamentally teams are not necessarily stronger on the road even if they are further up the table and that is where the first set of problems were arising
  4. Therefore, what if a team was so dominant in terms of league position difference but away from home?
  5. In this case, it would be a variation of step 1 but the difference would be they could either be better away from home OR be more than five places clear in the league table
  6. That way, a team could not necessarily be the best on the road but if say Tottenham are playing Aston Villa, with a 12 place gap between the two sides. Tottenham would still be favourites regardless of patchy away form
  7. While if certain criteria are not met in either step 1 or step 6, then the other option is a draw

With that in mind, lets see what ‘PremBot’ has in store for this split gameweek:

Step By Step

A list that to be honest does not look too illogical in all fairness (I guess I would say that) but with the step by step logic that has been laid out above, it certainly looks more intune with what I would pick in terms of gut feel.

Hopefully this is more the missing link to these data driven football predictions, as the past few weeks have certainly stuttered and therefore it will be interesting to see how many are correct over the course of this wekeend and next.

I’ll make this a V3 in my developer notes, so this will be the new base model going forward and this week will be a case of just seeing how much is correct. While if you like these picks, you are more than welcome to utilise them for yourself.

Happy punting and thanks for reading. Dan

If this has grabbed your interest and you would like to discuss/feedback then please feel free to drop me a message at While I am always looking for new football/data projects to work on and if you feel that my skills would be of use, I can be contacted at the same address.

Follow me on Facebook at Dan The Stat Man
You can also check out my Premier League Podcast on Soundcloud

Leave a Reply

Your email address will not be published. Required fields are marked *