We have spoken about how using goal and corners averages combined with football ratings systems can provide a powerful tool for forecasting football matches. Using some concepts that some of us may remember from school (if you can cast your mind back that far) we can start to predict the outcome of some aspects of a football match. This is a great addition to any football betting strategy. Lets look at how probability theory and statistics can help with correct score betting.
Why correct score football betting?
Correct scores as you might think are difficult to predict. If best a long shot... As such punters don't bet large amounts of money on correct scores. Becuase this isn't a focus for the bookies their fixed odds may offer value more often than you might think. The science and some extent the art is to spot these occasions.
Imagine odds of 109/1 on a top flight premier league team being beaten 2 - 0 at home. Using thise methods you would have seen value in those odds and gone in for the kill.
Well, Poisson (1781 - 1840) thought up a clever way of calculating the liklihood (or probability) of a number of events happening in a fixed time frame. Poisson distribution can be really useful for predicting correct scores in football games. Something a 19th Centuary French Mathematician could never have conceived in his wildest dreams.
Why is Poisson's Model helpful for predicting correct score results? Certain conditions are met to make it a good model to forecast correct score.
It calculates the probability of a number of random events occuring in a specific time set
There should be equal chance of the events occuring in each period
Thse events should be rate, i.e., not loads of them. So not so good for Basketball or American Football.
Sound familiar? You've guessed it....goals in a football match.
The elements that conincide in a football game to produce a goal are random. Also, goals don't happen ubiquitously in a match.
You can calculate the probability of an number of goals being scored during a match as long as you know the team's avereage (or mean) goal number. The are various websites suppying football data that can help with this.
We there is no need to route out your old maths books. It's a only a matter of entering the number into the Poisson equation and seeing what spits out. Luckily you can use the Poisson statistical function in an excell spreadsheet to do this. Just look at the help section to see where to enter the figures and what the end results means.
For the mathematical geeks amoung us you may have noticed a floor in this system. There isn't an equal chance of goals be scored by either team, becuase some teams are better than others. You wouldn't expect that Brazil has got and equal chance of scoring goals when playing Uruguay in the World Cup, now would you?
To get the most accurate results you must evaluate the relative strength of each team. We suggest using goal averages for the teams when the have played each other before (better than using goal averages from all games playing teams of varying competancy). It is best to select both home and away figures. It you don't have sufficient data use goal averages when the teams has played opponents of comparable competative strength. This is easy if you use the services such as those provided by footballbettingdata. Also, it is easy to obtain goal supremacy figures from sports spreadbetting companies. So you could could use this to take into account the relative strength of each team.
The Poisson distribution can give us the percentage probability of each side failing to score. Whats the chance of getting a 0-0 result? Our calculations tell us that there is a 20 and 30 percent chance of each team failing to score. To work out the combined liklihood we multiply the decimal equivalnet of the percentages together; .20 x .30 (=0.06). The result has a 6% chance of happening.
Typical probabilities may come in at 11.5% for a 1 -1 draw, or a 7.8% chance of a 1 - 0 away win. These percentages just show how unlikely you are to pick the exact score, which is reflected in the odd of course.
So, how do we use the percentage probability to find the best correct score odds?
First convert the percentage probability to decimal odds. You do this by dividing 1 by the decimal version of the percentage. An 11.5% chance of a 1 - 1 draw gives odds of 8.7 (1 / 0.115). Odds of 8.7 or more would be attractive. But, to take into account any error in our calculation and to build in a profit margin you may want to add 10%. If you add more you are less likli to find any value bets. In this example, you wouldn't except any bets unless they have decimal odds of 9.6 or above.
If you have not nodded off, you'll probably have figured out that you are going to have more loseing bets than winning ones with this football betting strategy. However, due to those large odds the winning bets will outweigh the loseing bets. Also this can be a very volatile market so you could have a bit of a rollacoster ride.
Don't forget the bookmakers have all the same information you do and more. However, their odds don't always reflect the true probabilities for various reasons. For example, offering longer odds to attract more punters to a bet or shortening odds to discourage punters.
With the aid of a betting exchange like Betfair you can become your own correct score bookmaker. Betfair allows you to lay correct scores. Using the same method. Estimate the true probability of a correct score. Instead of adding 10% deduct 10% from the decimal odds.
Now all you need to do is lay the result at your reduced odds and hope another member matches your bet. There we have it you are now the bookmaker.