cosmicsports
Colt
I think not, but just in case someone has an idea I 'm posting this.
What is it about ?
As we know the bookies first act as prognosticators and work out probabilities for the matches, then they convert these probabilities to prices and publish them in their websites for us to make bets. So for example with probabilities 50-30-20 and 5% bookie takeoff margin they 'd give us 1.90-3.17-4.75.
Now to what degree is their prognosis succseful ?
I mined some league data from the last five years from all over Europe and it comes out as 38.3%.
This is the mean value that comes out from the information sum.
Now consider this:
Suppose the match is one between team "A" and team "B".
I look into the last five matches of "A" and "B". The odds offered for those last five matches and the results.
Those odds can be converted back into probabilities of course.
Suppose then the "A" were according to the odds expected to have won 3 of their 5 matches but have won only 1.
Suppose also that "B" shows no deviation from what was expected.
So should n't that lower A's probability for the new match against "B" ?
I tried various formulations but the information sums stay immobile - rather can be made worse but not better.
It is possible of course the bookies have applied such fixes before delivering us the odds so I cannot do it again - I 'm only adding zero information to the system.
But initially I expected something better.
Yet sometimes our intuition discovers things.
So is there some hidden angle in this ?
What is it about ?
As we know the bookies first act as prognosticators and work out probabilities for the matches, then they convert these probabilities to prices and publish them in their websites for us to make bets. So for example with probabilities 50-30-20 and 5% bookie takeoff margin they 'd give us 1.90-3.17-4.75.
Now to what degree is their prognosis succseful ?
I mined some league data from the last five years from all over Europe and it comes out as 38.3%.
This is the mean value that comes out from the information sum.
Now consider this:
Suppose the match is one between team "A" and team "B".
I look into the last five matches of "A" and "B". The odds offered for those last five matches and the results.
Those odds can be converted back into probabilities of course.
Suppose then the "A" were according to the odds expected to have won 3 of their 5 matches but have won only 1.
Suppose also that "B" shows no deviation from what was expected.
So should n't that lower A's probability for the new match against "B" ?
I tried various formulations but the information sums stay immobile - rather can be made worse but not better.
It is possible of course the bookies have applied such fixes before delivering us the odds so I cannot do it again - I 'm only adding zero information to the system.
But initially I expected something better.
Yet sometimes our intuition discovers things.
So is there some hidden angle in this ?
Last edited: