How it works

If the real probability of the event turned out to be 20%, then for 100 bets we will get 20 wins. By making equal bets at $100, we will bet 100*100 = $10,000, and win 100*20*4 = $8,000. Thus, the bookmaker will make a positive gain of $2,000.

If the real probability of the event turned out to be 30%, then for 100 bets we get 30 wins. Therefore, with the same bets, we will win 100*30*4 = $12,000 and now we will be in the black with $2,000.

The bookmaker determines that the probability of a win for each tennis player is 50%. This is equivalent to odds of 2.00 against 2.00 At the opening of the betting line, the bookmaker sets the odds taking into account his margin of 5% - 1.90 against 1.90. If 100 people place $100 on each side, the bookmaker will gain 100 x 100 x 2 = $20,000. From one side’s win he needs to pay out the winners 100 x 100 x 1.9 = $19,000. Thus, net profit will be $1,000.

Suppose, for some reason, most players are confident in the victory of the first tennis player and bets are distributed according to the proportion: 80% (160 people) to 20% (40 people). In this case, if the bookmaker leaves the odds based on probability, in case the first tennis player wins, he will have to pay 150 x 100 x 1.9 = $ 30,400 and take a loss of $10,400. Therefore, it is more profitable for a bookmaker to change the odds in accordance with the market, rather than the probability of the event. The odds corresponding to 80% and 20% are 1.25 and 5.00. With a margin of 5%, this will be 1.20 and 4.6. Then, if the first tennis player wins, the bookmaker will need to pay 160 x 100 x 1.2 = $19,200. And with the victory of the second 40 x 100 x 4.6 = $18,400. Thus, in both cases, the payout will be less than the collected amount of $20,000, which will allow the bookmaker in any case to be in the black.

The score of a football match in which one team was perceived as the clear favorite is still a draw. There are 10 minutes left until the end of the game. The favorite’s odds of winning are 4.60 against 1.20 on the outcome that they won’t win. Suddenly, the underdog receives a red card and the favorite team is in the majority. The odds change to 4.00 against 1.25. For most players, this often turns out to be a signal for action and they cause the fall of the odds for the favorite to 3.50, raising the quote for the underdog up to 1.30.

For most of them, it does not matter which odds the bookmaker gives. Logic is limited by the mindset “the favorite is in the majority, so they will pull off a win”. But in reality, the favorite is not always pressed to win. Often, playing in the minority only provokes the underdog to deeply defend a draw, playing to keep the score.

Of course, after receiving a red card, the probability of the underdog saving the draw falls. But it falls by only 3%, while the odds fall by 6% due to the market's reaction to the bright marker. Due to the full action in live play, the bookmaker has to use automatic tools to adjust the odds. Meanwhile, our algorithm determines what amount of bets for this situation are abnormal and at what stage the odds have become overvalued.

Our algorithm analyzes hundreds of match events: cards, corners, fouls, leadership of the home team, the guests, the percentage of possession and many markers that are not obvious to the human eye, which in practice have been identified by the neural network as important. Next, we get the artificial neural network to look for different combinations between these events and their variations. The ANN saves those patterns in the game that players react to and tries to determine the deviation rate from the current odds when such a pattern appears, comparing the results with new data and making corrections based on the new data. Thus, the network is self-learning.

Those algorithms that show favorable results over the long term are separated out by the program into a table in the form of rates. Rates and their results are saved and then the program compares the data and uses it for self-training.

The program is constantly being developed by the Bethub team. It is populated with more data for analysis, and we connect new markers and tools to it. As a result, despite the diminishing volatility of the betting lines and the high pace of technical development of the bookmakers themselves, our software is becoming more stable, stronger and widespread each year.

The maximum acceptable drop in odds is 3%. This value is the average revaluation of the odds by a bookmaker based on our market research and software statistics.

If you lose even 1%, then over a long term you can incur a loss instead of a profit.

Remember that the tool is free and is in the public domain. Thousands of players use it every day. Therefore, catching the latest odds is not always easy.

In extreme cases, the number of bets on each event should be fixed. But in this case, you need to limit yourself only to big markets.

The fact is that the algorithm determines only the vector of the revalued market. 25-40% of matches may be incorrectly determined. This is a normal indicator in which the overall result remains positive. But if you place bets unevenly, then most of them may end up in incorrectly defined matches.

Unfortunately, we do not know what is happening inside the neural network. We don’t know which bets were determined correctly and which odds turned out to be overvalued (if you figure out how to find out, feel free to apply for a Turing award). Therefore, it is very important to make equal and stable bets.

The drawdown of the program on low-profit bookmakers (with a large margin) during the tests reached 75%. Thus, only rates not exceeding 1.3% of the jackpot could guarantee the safety of the jackpot through the entire time of testing the software. There is also no guarantee that this was the maximum drawdown level, despite statistics of 100,000 bets. Also, with mass use, there is always the possibility of mistakes, both technical on our part and due to the human factor on the part of the players.

It is very difficult to make big profits on public tools used by the masses. The success story of all professional players is in the large body of their developments and in the algorithms that appear gradually as you study the various disciplines related to betting.

Now you have a huge listing of currency rates, which works 7 days a week around the clock in automatic mode. Constantly seeing revalued odds in the betting line, you learn to understand the market, its movement and the actions of players.

Our research of the bets by players who used our software showed that even by simply duplicating 100 bets every day for several months, the players improved their performance, making further bets independently in comparison with the documented results before using our software.