A Profitable Model For Predicting the Over/Under Market in Football

[Raskol(nee Kov)]

Ok I did some using stats for the last 12 games for each team and this is my small selection just for fun. Curiously they are all saying to Back Under 2.5 (or lay overs). Did anyone find any where it suggested otherwise to back overs 2.5?


Over Under 2.5 goals Sample Test 270226.png

Which site did you use to get that data?

[gstar1975]

Ok I did some using stats for the last 12 games for each team and this is my small selection just for fun. Curiously they are all saying to Back Under 2.5 (or lay overs). Did anyone find any where it suggested otherwise to back overs 2.5?


Over Under 2.5 goals Sample Test 270226.png

Which site did you use to get that data?

From thestatsdontlie for the shots but then realised you have to pay for the corner data. So I asked Gemini and used that data. Not sure if its correct. I will search for another site.

[gstar1975]

There is also the question of timing, when to place bets. Also could this infomation be used to take a position inplay?

[Bet79]

Looking at football data is it case of adding up the last 8-12 games worth of ‘HS’ and finding the average to enter into cell C3?

[gstar1975]

Looking at football data is it case of adding up the last 8-12 games worth of ‘HS’ and finding the average to enter into cell C3?

Yes.

[Bet79]

Thanks

[Vanila]

I would like a step by step explanation sir on this one

[Mike Oxlong]

Can’t help but wonder why they didn’t use expected goals in their calculations? I understand they stated that they wanted to use ‘free data on overall match statistics’ and avoided paying for further data, but is there not free sources of xG data? Admittedly, I have not done any looking myself.

I suppose the recent Opta contract termination with FBref removed the data from there, but this paper was published in 2020.

[Filbal]

I found this paper on-line and had a look.

https://researchonline.lse.ac.uk/id/epr ... ls_LSE.pdf

In this video, I explain roughly what they are doing and the key formula: -

https://youtu.be/CK7LyOODP6c

I've created a spreadsheet that replicates this, which I have attached.

Peter, I have a doubt. The odds we should use as a reference to place our bets on the exchange, should they be the ones that come from the pressure model averaged 70% with the market, or the ones that come from the logistic regression? Just to be clear, should we take as reference the one I circled in black or the one I circled in red?