Compiling Soccer Odds for Dummies

Why would anyone want to compile their own odds? After all, betting on soccer is about finding winners and you reckon you are pretty good at that. Wrong!

"Any fool should be able to achieve an 80% plus strike rate but making a profit from these bets is another matter."

Any fool should be able to achieve an 80% plus strike rate but making a profit from these bets is another matter. It is assumed that you already understand why betting 10/11 about a coin toss game will lead to long term losses and that probability is expressed by a number between 0 (will not happen) and 1 (certain to happen). If not visit Peter Webb's excellent site and get clued up first.

We have no chance of profiting from football betting unless we can identify and take odds that are in our favour, commonly known as “Value”. 11/10 about a coin toss is obviously value as the true odds are evens or a 50% chance. To get value from soccer betting we must be able to identify when value exists and to do that we need to compile our own odds for the matches.

"To get value from soccer betting we must be able to identify when value exists and to do that we need to compile our own odds for the matches."

The purpose of this article is not to provide a foolproof method of odds compiling (if only one existed) but to provide food for thought to inspire the novice to start thinking about how he may tackle the problem in his own way. It should be understood that none of the methods suggested are in any way sophisticated enough to be used in anger.

Head to Heads.

Given a sufficiently large sample we could use the head-to-head record of teams to compute a rough estimate of the odds for the game.

Let’s take the Hibernian v Celtic game of 17th September. The head-to-head stats for that game going back 15 years are home win 5, draw 8 and away win 18; a total of 31 games. We can express the probability of a home win as 5/31 or 0.16. To convert probability to decimal odds we invert, that is we divide 1 by the probability. Thus the estimated decimal odds for the home win are given by 1/0.16 = 6.2. In Excel we would set it up as below, needing only to input the number of games in the yellow cells.

The last row gives the actual bookmakers odds for the game and as can be seen the estimates are not a million miles away from the offered odds. That should not surprise of course as the example was carefully chosen. In reality there are insufficient recent h-to-h data available to make this a viable method.

In this brief introduction we have seen how form data can be easily converted to estimated odds for a game. In the next article we will look at some realistic methods.