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 44 (from October 31 to November 6 of 2022).

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.

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

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

As we can see the highest ROI_O on this week is for strategy A-B and the lowest ROI_O is for strategy A+B. The ROI_U on this week is -100% for all strategies. The highest ROI_T for all bets on this week is for strategy A-B and the lowest ROI_T is for strategy A+B.

Total ROIs per each week are shown in the table below.

If we would invested equal sums into these bets every week, our ROI (on the total investment) per four weeks would be 24.12% for strategy A-B (it is equivalent to over 250% annually) and 10.29% for strategy A.

The additional advantage of this type of investment (compare with other types) is that you do not risk the whole invested capital, only a fraction of it. Here is an example. Suppose that each week you invest $1000 into bets (10 bets with $100 per each bet). The total invested capital (for four weeks) is $4000, but capital at risk (per each week) is only $1000. In fact, the capital at risk is lower than $1000 because your bets per game are less than $1000. With strategy A-B you would have $960 of income per four weeks from this activity.

The next advantage of this type of investment is that it does not correlate with financial markets and can not be manipulated by central banks and governments.

An investor can think of investing in bets as investments in high risk NFTs or DeFi projects with high ROIs. The difference only that instead of locking invested capital on many months, it is locked only on a couple of hours (duration of soccer games). This allows investors to increase speed of generation of positive returns on small amounts of available capital.

For the reason that investments are diversified over many bets there is no a concentration of risks in this type of investing. Results of games are not correlated with each other, therefore risks of high loses per day are low.

Another advantage for crypto investors is an opportunity to use blockchain based companies, some of which do not have KYC requirements. Here are some of them:

Most likely some of them will be forced to implement KYC requirements in the near future, but it is possible that some will allow anonymous betting.

For calculations of ROIs we use odds from these sources:

https://www.football-data.co.uk

The previous post https://www.publish0x.com/analysis-of-predictions-in-sports/analysis-of-predictions-in-sports-week-43-xpzllgv

P.S.

On these blogs you can find solutions to many problems:

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

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