If you are interested from computer generated betting predictions, please have a look on this site - Basketball Predictor. So far it has great results with NBA matches.

Computer made Under Over soccer predictions

One of the most interesting betting options on football matches is whether both teams will score more or less than 2.5 goals in the game. Of course, this bet has many variations and can be played on 0.5, 1.5, 3.5 and more goals, the goal only for one the two teams or only these scored in one of the halves. There are many variations, but the hard task is to make profit from all of these kind of bets.

Is it possible to create a computer program that provides fairly true betting predictions for the number of the goals scored in a match? The answer to this question hasn’t been answered so far or rather it goes to one big No, but recently I came across a curious trial in this area. I found the site 1f6.info and since the latest predictions is from the European Championship in 2016 we can conclude that the experiment was unsuccessful.

However, the approach of the developers is interesting and deserves attention. Here is briefly what they most probably were trying to do and how far they reached.

The idea of ​​this computer program is to estimate the number of goals scored in a football match by calculating the number of created chances to score by both teams and the number of actually scored goals. For base are taken the last few games of each of the teams and then the program calculates the percentage probability of each one of the teams to score. After having these number the program makes projections which simulate a number of matches between the two teams to get the percentage probability of how much would the total number of goals scored in the game be. This probability is assessed against the odds and where it found a value a bet must be made.

There is one additional condition and it is that only matches with probabilities above 80 percent and odds (taken from Unibet sports betting) more than 1.4 are considered for real betting. Thus guarantee sufficient security of the made investments.

What are the results? In fact, it wasn’t very easy to take out the final results because such were not represented on the site, but after some calculations, it appears that this approach returns very little profit after 100 bets. Presumably, that is why there are no recent publications on the site.

Anyway, the idea is interesting and definitely worth consideration. At least, this approach to sports betting is something that very few use but it has to be the opposite.

Both comments and pings are currently closed.