6. More About Reliabilities

Anybody who understands the basics of football can make a prediction for the outcome of an upcoming match. However, the validity of any such prediction would generally be uncertain to a greater or lesser degree, according to which teams are involved. That can best be demonstrated by considering what would happen if you were to bet on all the European Premier League football matches in any given football season turning out to be Home Wins. If you were to research past records about this, you would see that only 45% of such bets would prove to be correct. Of course, the Bookies have the Odds for that ‘blanket betting method’ adequately covered, so you would only lose money if betting in that manner.

Further, even if you were to apply in-depth analytical techniques of the performance rating of the teams (to try to identify more precisely which matches would most likely be in that list of 45% actual Home Wins), it would still be hard to get more than 50% of them right. The fact that you use perfectly sound logic to determine the anticipated match outcomes would make no difference whatsoever. That is because there is a large amount of randomness evident when it comes to the performance of all but the top-most football teams. And, even there, the best (strongest) teams will still let you down; that is simply the nature of the Beautiful Game!

This therefore means that, in order to make good betting decisions, you will need to take cognisance of the relative combined ‘Reliabilities’ of the two teams in a match to deliver their average level of performance. Only in that way would you be able to select the best matches to bet on and have a chance to determine which matches it would (perhaps) be best to ignore. In fact, the importance of looking into and understanding the range of different Reliabilities that our Program measures cannot be stressed enough. If we did not use our Program’s Reliabilities information adequately, we know from experience that the list of matches derived for betting purposes would be no more reliable than if we had rolled dice to determine what those matches should be.

Our Program attempts not only to predict the match result (Home Win, Away Win or Draw) but also the most likely score-line for each match. As we mentioned earlier, that is all done based on algorithms centred around the ‘Law of Averages’. However, since that Law is a generalised application that often does not reflect reality in a specific given situation/setting, it is evident that the Reliabilities of any predictions applying that Law need to be assessed very carefully. For those interested in such things, please note that we found that the modified (some say adulterated) view of the ‘Pareto Principle’ (i.e. 80% of what transpires results from 20% of the causal factors) cannot be applied to the databases we utilise within our Program and certainly not for assessing the most likely score-lines for the matches, primarily because the spread of score-lines achieved in any given season is extremely wide for most teams (almost unbelievably so).

After having decided which Prediction Method is to be used as the basis for conducting the Selections Process, the next step in our Program’s Selections Process requires the User to decide what sortation factors are to be used for pulling through the most likely winning predictions from amongst the many matches that take place in the Premier European Football Divisions each week (on average, around 150 matches per week). According to what type of betting is being considered, the choice of the factors to be employed and their relative significance to other factors will inevitably vary. A great deal of thought needs to be put into this particular decision-making process, and our Program provides many different reporting outputs to help the User decide what the best factors to employ for match selection purposes should be.

Since it would be very risky to place reliance on just one single Selection Parameter for making betting decisions, our Program makes it possible to incorporate up to 7 different lines of ‘Sortation Parameters’ in order to obtain a more balanced output for assessing the relative trustworthiness of the final selections produced.

[148/1000 to date] [for bonus subscription - see foot of Home screen]