Whether or not the usage of head to head statistics ( HTH's ) are of paramount importance to you when choosing a given pick, the fact remains that the usage of these statistics is deeply ingrained in the psyche of many of the on-line soccer betting community.

However, are we correct to place such emphasis on HTH's ? Is there really any correlation between a team having a good record against their opponents in previous games and an increased likelihood of them winning their next encounter? Footy4cast upsets the wife once more to bring you the results of literally hours of study to bring you the following analysis.

A Long Journey - But we got There in the End ?!

To assess whether their was a positive correlation between good HTH's and a high percentage of wins in the next game, Footy4cast looked at all the HTH records of all the teams in six major European leagues ( English Premier, Spanish LA Liga, Italian Serie A, French Le Campionnat and Portuguese 1 ).. The only stipulation was that the home team had hosted their opponents at least 6 times between 1992 and 2001/02. ( Please note that this study is only in relation to the HTH's stats for home teams ).

To gauge performance of the home teams against their various opposition, a % of points was calculated in the following format. For instance :-

Man Utd v Aston Villa ( example only ) = WWWWDLWWWD. The total points that Man Utd acquired from these 10 games ( 1992 to 2001/02 ) was 23. This was then divided by 10 to give an average total of 2.30. A % HTH performance rating was then found by dividing 2.30 by 3 to give 76.66%.. This percentage figure was acquired for the home HTH performance of every combination of teams in each division.

Once this task was completed ( and my wife filed for divorce !!! ) all the matches were grouped according to the following ranges of performance 00-20%, 21-40%,41-60%,61-80% and 81-100%. Footy4cast then looked at the result of all the matches within these ranges in the 2002/03 season, to see how well the home team performed. Obviously, if there is justification for using HTH's in our selection criteria, then we should see high percentages of home wins in 2002/03 for head to matches which showed high HTH performance ratings ( between 1992 & 2001/02 ).

The Results

Whilst the results are some way off producing a positive correlation which would be justifiable in relation to a scientific experiment, they do show the following :-

Home H2H Percentage

As you can see from the results above, the percentage of home wins achieved in 2002/03 definitely rises as the HTH performance ratings (ranges) also rises. Surely if head to heads should be excluded entirely from our selection criteria then there would be absolutely no steady rises in 2002/03 home win percentages - head to heads then IMO are a must for inclusion in your selection criteria.

So are there any Potential Systems?

Footy4cast went one step further and back tested all the games within the various ranges to level stakes ( using William Hill Odds as these are generally one of the lowest on offer ) to see whether we could have made money just by considering HTH performances from 1992 - 2001/02 to make home picks for the 2002/03.

The yields were as follows :-

00-20% = -05.65%
21-40% = -11.69%
41-60% = -19.23%
61-80% = -08.89%
81-100% = +08.88%

As you can see backing all teams in 2002/03 who had achieved an 81-100% home record against their various opposition between 1992-2001/03, would have produced a highly profitable +08.88% yield to level stakes ( in exactly one hundred selections up to 14/04/03 )…for doing absolutely nothing except go through every HTH in all 6 leagues prior to the start of each domestic season.

Have you got time to do this?

If not try and make time !!!

You might be presently surprised !?

And Finally...

Here I have to hold my hands up a little bit. This is a system I have employed for several seasons with varying degrees of success, however the lowest yield I have had since 99/00 is 05.65%.

In my opinion the bookies leave themselves open with top range head to heads ( i.e 81-100% ) as they still have to take into account form even when a home team has perfect head to heads.

This means that average odds for the picks usually level out at around 1.55-1.62 meaning that a 74% winning success rate for the home teams in the following season is often exploitable to level stakes.

I should also point out that I use a range of 17 preferred bookies which means that I have maximised my yield. This season has paid me 11.01% by taking best odds on offer.

The problem with the system is that the matches to be included change constantly, from season to season, as I only ever use a decade of HTH's - a match that may be applicable this season in the 81-100% range could change next season and fall out of this range.

So the matches need to be assessed in the close season every year, which is massively time consuming !!!

Many thanks for your attention and I hope this article has been of some interest.

References

www.nifs.no