Posted:
#1
Looking for some ways to make money on the over / under bets I thought I might use the Poisson function in the Excel worksheet to compute the real odds for games to go under or over.
First of all I should say that Poisson distribution only applies to soccer and hockey games where scoring is a Poisson process.
The Poisson distribution of probabilities can be used to estimate the probabilities of occurence for events that happen one at a time, are relatively rare, we know what's their frequency of occurence in a given time interval and their occurence is independendent one of each other.
The goals in soccer and hockey games are such events. They are relatively rare, have manny opportunities to occur (there are a lot of shots on goal during a game but only 5-6 of them become goals) and they happen only one at a time. In football, baseball and basketball scoring is not a Poisson process, because the points may and do occur 2, 3 or more at a given time.
Anyway, it has been established that scoring in hockey is a Poisson process. The goals are randomly distributed over the 3 periods with some exceptions that can be ignored.
I thought that hockey being an offensive sport I should use only the average of goals scored by a team. In hockey goals are much more frequent than in soccer and in order to win you have to score one more than your opponent does.
In soccer you can win points by defending the draw so there are a lot of team who developed very defensive strategies and I think that prediction must take into consideration the defense as well as the offense.
But I mainly focus my strategy on hockey games, so that's how I thought predicting the hockey total:
Example from today's Russian KHL games:
Voskreshensk plays at home against Novosibirsk. Voskreshensk scores at home 1.72 goals / game and Novosibirsk scores away 1.95 goals / game. The average occurence of goals / game in this situation would be 3.67 goals / game (???, that's it, they're very low scoring). I compute the odds for under 5.5 using the worksheet in Excel and I find that chances are 83.4% for the total goals to be under 5.5. I go to Ladbrokes, my bookie, who, for my surprise, has the total listed at 5.5 and decimal odds of 1.6 (that means a value of 40%) and I place the bet (no Kelly criterion, just flat betting, one unit / play) and hope that I won't fall in the 166 / 1000 opportunities for this game should go over 5.5 accordingly to Poisson distribution of probabilities.
Using the scoring avg. of each team I compute the chances of being scored 0,1,2,3,4 and 5 goals and then, by adding these I obtain the chances for under 5.5. Then 1-chances of under 5.5 gives me the chances to go over 5.5.
I set my rules to not take plays with chances under 60% and pay attention to their value. They must have just a little value in order to play them.
I am curiously to see if I will reach a 60% hit rate, using these rules.
The picks for today's hockey games would be (in european leagues):
1. Voskreshensk - Novosibirsk under 5.5 - 83%
2. Cherepovec - TVER THK under 5.5 - 67%
3. Karpat - IFK Helsinki over 5.5 - 64%
These are all plays that have > 60% chances and a little value at my bookie's prices.
I'll come back later with the results.
I'm also curious if you tried it that way and what results did you get.
0
Hot Forum Topics
