Most football match prediction use a points rating scheme. The number of points which a team is awarded is dependent upon the outcome of previous matches involving that team. More points are awarded for previous match victories than for draws or losses. For example win 4 points, draw 1 and lose 0 points. Sometimes points are awarded on the basis of other match statistics, most usually goals scored and conceded, but also shots on goal or any other data . A higher team points rating means a greater team winning chance.
By comparing the winning potential of opposing teams we can calculate a model for possible outcome of the match. Rather than making a specific prediction, however, the outcome of a match is expressed in the form of a probability distribution. Consider, for example, a match between Levski Sofia. (home) and CSKA (away). Using the described methodology, we might calculate the following probability distribution for the three possible match outcomes:
Levski Sofia. vs. CSKA: Home 55% Draw 25% Away 20%
On its own, this probability distribution does not tell us very much other than the most likely outcome of this match is a Home Win for Levski Sofia.
However, given that the average probability distribution of league match outcomes is 46% home win, 27% draw, 27% away win (for Bulgarian league), it is clear that betting on a Home team victory is the most likely outcome. To utilise this match information in a meaningful way, we need to have a look at what betting odds are being offered by the bookmaker.
Premiership
Primera Division
Serie A
Germany
France
Scotland
Eredivisie
Jupiler League
Superliga
T-Mob. Bundesliga
Super Ligi
Super League
SAS Ligaen
Hi!
I’ve been using a system based on GD (similar to yours) for the last 2 mounths… and for the first time in 2 years I’m getting 7/8 correct results (1X2) out of each ten, and even correct scores (2/3 in 10).
The only difference is that I’m focusing on the last 8 games of each team form, and I also added a very important feature to my predictions: Record Versus Top-Half Opposition and Versus Bottom-Half Opposition.
I’ve notest that my predictions are very similar to the ones of prosoccer, and that’s because of the “Poisson Distribution Calculator” that I use to get the final predictions.
Altough, I don’t understand how (if in fact I understood what’s written in the first paragraph of the text above) you use any other kind of data (shots on goal, for example, as mentioned)? If that’s the case, excellent work.
All in all, I want to congratulate the creators of this site for the innovative way of predicting the outcome of a football match, and I wish all the best to all of you.
Hope to see more developments soon.
p.s.: sorry for my bad english
Last version of the system use Poisson distribution as base. We found that both systems “Poisson” and “Goal difference” give almost the same results for last 5-6 games.
I’m going to calibrate the way system choose final picks. This is scheduled for first weeks of the new year.
By the way, for my own purpose i create this page http://www.zaloji.net .It use database from pro-soccer.info, but calculate picks based on value odds.
cheers
Football is a bit tiring game but it has given me great excitement and it is a good physical exercise too.”-`