Same As Before

Same As Before

If you have come across the results page for this project, you will be aware that the first week of live testing for my English Premier League model saw a strike rate of 60% or six out ten correct predictions in layman terms.

A return which can be considered to be just on the fringes of the Europa League places and although it didn’t lead me to an embarrassment of riches, it certainly was not the worst start to proceedings either.

Therefore with a decent grounding in all things predictive, it is time to recap on the picks for the most recent weekend of action:

Same As Before

And just like last week, I once again returned six out of ten picks from my predictive model and that keeps the burgeoning strike rate at a respectable 60% – again not bad, but we are going to have to do a little better over the course of the next few weeks


While although I got six out of ten picks correct, it was one of those weeks where it either could have been a lot worse or a lot better – for example, you only have to look at the late winners that Manchester City and Liverpool scored for testament of that.

More importantly, I could have built 1,000 variations of a predictive model and none of them would have seen either an Aston Villa or Southampton win, so it just goes to show that shocks are going to be a huge and arguably unavoidable factor in all of this.

I’m not going to bemoan every pick that I didn’t get, because otherwise it will make this project rather futile, although I will have to eat an element of humble pie after making a statement in regard to the picks that were made on Saturday morning.

I said that there were three red flags this weekend, a trio that looked like (- equals my gut feel pick at the start of the weekend)

Bournemouth vs Manchester United – Manchester United Win
Arsenal vs Wolves – Arsenal Win
West Ham vs Newcastle – West Ham Win

I thought that the Gunners would win at home on Saturday but for the second week in a row, Unai Emery’s men were held at the Emirates and for the second week in a row, the model was smarter than me.

Which means that preconceived bias is a dangerous factor, because I’ve made the crucial mistake (from a gut instinct) of putting too much expectation in a big six club (the same for a Manchester United win) and therefore need to trust my model more than any prior conceptions I may have.

However, the model let me down on both the Vitality and London Stadium fronts as the victors were the opposite to my thoughts above and more importantly they were still a mile apart from what my model predicted.

That said, six out of ten is not to be sniffed at and it does at least point to the fact that there is some mileage in this. Then again, it does not need football’s biggest brains to predict a Liverpool or Manchester City win right now.

Which means it is going to be really interesting when the two teams meet at Anfield next Sunday, I don’t know what the model is going to predict for that one (well at least not yet) and once again after a relatively decent performance, it does beg the question as to whether or not changes should be made.

That’s something that I will deliberate in the week but once again the notion of backing the ‘big six’ seems to have been blown out of the water and although this makes life a little harder, it certainly makes it more interesting at the same time.

I’m now going to go away and tinker with the model (it needs a name really), while if anyone has any suggestions I am all ears (feedback is always welcome) and I’ll report back with my findings at some point next week.

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 *