Don't Bet Invest - Efficiency Calculations Explained

[Vince A]

 

Hello guys. 

So I've been using this system for awhile now @Dontbetinvest. 

I found their algorithms to be much better than the market's efficiencies and values, with their true odds helping me analyze better my bets. 

I hope it helps you all, and below is a deep explanation on the entire system.

The DBinvest Efficiency table shows the history of how DB invest’s true percentages are faring compared to theoretical probabilities. In other words, it demonstrates whether our algorithm is efficient or not. To make it practical, we divided all of DBI’s calculated true probabilities (where their respective bets carry value over 0%) into groups of 10% (cluster groups or % range). We are going to explain the above table with the example below.

Let us say that Barcelona’s true winning percentage versus a specific opponent is 57.25% (calculated automatically by our algorithm). This percentage belongs to the 50%-60% cluster group (percentage range).

Barcelona won. Now we have 25979 won games out of 48607 total bets in the same cluster group.

By dividing 25979/48607 we get an actual hit rate of 53.45%. In other words, if we place 100 bets that range between the 50-60% true percentages, we are going to end up winning 53.45 times. Let’s round it up to 53 times.

In order to measure our cluster’s actual hit rate and see how good or bad it is, we have to compare it with the theoretical hit rate of the same cluster group. To do so, we add the two ranges divided by 2. So, (50+60)/2 = 55%. Therefore, theoretically speaking, if we buy 100 bets that have true percentages between 50 and 60%, we should be winning 55 times.

By dividing the actual hit rate over the theoretical hit rate of this 50-60% cluster times 100, we get a clearer picture of how our cluster group is faring compared to the perfect half: 55% (the theoretical hit rate):
(53.45/55)*100 = 97.18% (what we call the value coefficient).

This means we are 2.82% away from achieving perfect calculations in this specific cluster group, or to put it better, we are 97.18% accurate. On the other hand, we managed to exceed the theoretical probabilities in other cluster groups, like the 0-10% and 80-90% clusters... this proves how accurate our algorithm really is. Simply put, the more bets we record, the closer to perfection our algorithm becomes.

Here is where you can now conclude DB invest’s true probabilities and prices and assess how efficient they are.