Analysis of predictions in sports: week 41


In this post we analyze how accurate are predictions on total goals over/under 2.5 for English Premier League games and how profitable are strategies, which are based on these predictions. We use predictions from source A https://www.ispreport.xyz/utools/predict/cur/predict.html

and source B https://www.predictz.com/predictions/today/over-under-25-goals/ , which are available for free.

 

We calculate accuracy and returns on investments for five strategies:

-strategy A, bet on predictions given by A;

-strategy B, bet on predictions given by B;

-strategy A+B, bet on predictions given by A and B if they are the same;

-strategy A-B, bet on predictions given by A if predictions of B are the opposite to A;

-strategy B-A, bet on predictions given by B if predictions of A are the opposite to B.

 

There are three types of accuracy:

-accuracy of over (ratio in % of accurate predictions of over to total number of predictions of over), we use AO for this type of accuracy;

-accuracy of under (ratio in % of accurate predictions of under to total number of predictions of under), we use AU for this type of accuracy;

-total accuracy (ratio in % of all accurate predictions to total number of all predictions), we use AT for this type of accuracy;

 

The table below shows all predictions and results for week 41 (from October 10 to 16 of 2022).

 

t1

We use “o” for over, “u” for under, and “n” for not to bet in this table.

 

The table below shows AO, AU, and AT (in %) for five strategies: A, B, A+B, A-B, B-A.

 

t2

As we can see the most accurate strategy for bets over on this week is A-B and the least accurate strategies are B, A+B, B-A. The most accurate strategies for bets under on this week are A, A+B, A-B and the least accurate is B-A. The most accurate strategy for all bets on this week is A+B and the least accurate is B.

 

The table below shows returns on investments (ROI_O, ROI_U, and ROI_T) for the five strategies: A, B, A+B, A-B, B-A in %.

 

t3

As we can see the highest ROI_O on this week is for strategy A-B and the lowest ROI_O is for strategies B, A+B, B-A. The highest ROI_U on this week is for strategy A-B and the lowest ROI_U is for strategy B-A. The highest ROI_T for all bets on this week is for strategy A+B and the lowest ROI_T is for strategy B.

 

The interesting observation is that by combining predictions A, B the total ROI more than doubled (from 4.1% in A to 10.6% in A+B). It will be interesting to see if this observation still will be true in the long run.

 

 

For calculations of ROIs we use odds from these sources: https://www.football-data.co.uk https://www.oddsportal.com

 

 

 

P.S.

 

On these blogs you can find solutions to many problems:

 

https://www.publish0x.com/practical-data-science-and-optimization-in-example

 

https://www.publish0x.com/simple-solutions-to-complex-problems

 

 

How do you rate this article?

13


I_g_o_r
I_g_o_r

I am curious about science, technologies and their applications to solving real problems.


Analysis of predictions in sports
Analysis of predictions in sports

On this blog you find posts with analysis of predictions in sports. The main focus of the analysis is on accuracy of predictions and profitability of betting strategies, which are based on these predictions.

Send a $0.01 microtip in crypto to the author, and earn yourself as you read!

20% to author / 80% to me.
We pay the tips from our rewards pool.