Football Team Payback - How to identify value

[Vilamoura04]

Team payback. Do you think this is a dependable method for identifying teams who are value to back, or is it a good way of identifying overachieving teams who are value to oppose? Team payback is how profitable a specified team is if you had backed it to win in every match for level stakes over a specified time period. For example, if you had backed Man United to win every game this season. Often this is split between home and away performance, so you would look at level stakes performance backing Man U in their home matches only. Team payback is often cited as a method for identifying teams who are unfashionable and therefore underestimated by the betting public. Bookmakers are generally thought to set odds to evenly split the bets they recieve from the betting public, rather than to evenly split the actual result. Therefore these teams unfashionable teams would be offered at value prices and are good value back propositions. But is this what it actually signifiies? Or does it signify teams who the bookmakers are pricing correctly, it is just that the team has been overachieving against their "true level" as indicated by the bookmaker prices. They can therefore be expected to "regress to the mean", or rather, sink back to their true level of performance in line with bookmaker and public expectations. Thus these are teams that, due to their overachievement, are probably now being offered at lower prices than their "true level" and therefore the value is actually in opposing, or lay these teams.

[Machine]

Re: Football Team Payback - How to identify value

Team payback. Do you think this is a dependable method for identifying teams who are value to back' date= or is it a good way of identifying overachieving teams who are value to oppose?

Yes depending on how you use it.

Team payback. Do you think this is a dependable method for identifying teams who are value to back, or is it a good way of identifying overachieving teams who are value to oppose? Team payback is how profitable a specified team is if you had backed it to win in every match for level stakes over a specified time period. For example, if you had backed Man United to win every game this season. Often this is split between home and away performance, so you would look at level stakes performance backing Man U in their home matches only.

Not exactly correct, due to the variance in the odds it is better to vary the stake according to the odds to return a fixed amount.

Or does it signify teams who the bookmakers are pricing correctly' date=' it is just that the team has been overachieving against their [b']"true level" as indicated by the bookmaker prices. They can therefore be expected to "regress to the mean", or rather, sink back to their true level of performance in line with bookmaker and public expectations.

Again not exactly the correct asumption. To assume the odds are normally distributed and to take the 'average/mean' of the odds is wrong in most cases. A simple test ...If mean=median=mode then you have normal distribution, if not you have to check for skewness Lets take your example Man United @ home [TABLE=width: 1086] [TR] [TD]E0[/TD] [TD=align: right]25/08/2012[/TD] [TD]Man United[/TD] [TD]Fulham[/TD] [TD=align: right]1.3[/TD] [TD]H[/TD] [TD][/TD] [TD=colspan: 2, align: center]Man United total home odds[/TD] [TD][/TD] [TD=colspan: 2, align: center]Man United total home win odds[/TD] [TD][/TD] [/TR] [TR] [TD]E0[/TD] [TD=align: right]15/09/2012[/TD] [TD]Man United[/TD] [TD]Wigan[/TD] [TD=align: right]1.25[/TD] [TD]H[/TD] [TD][/TD] [TD][/TD] [TD][/TD] [TD][/TD] [TD][/TD] [TD][/TD] [TD][/TD] [/TR] [TR] [TD]E0[/TD] [TD=align: right]29/09/2012[/TD] [TD]Man United[/TD] [TD]Tottenham[/TD] [TD=align: right]1.57[/TD] [TD]A[/TD] [TD][/TD] [TD]Mean[/TD] [TD=align: right]1.39[/TD] [TD][/TD] [TD]Mean[/TD] [TD=align: right]1.37[/TD] [TD][/TD] [/TR] [TR] [TD]E0[/TD] [TD=align: right]20/10/2012[/TD] [TD]Man United[/TD] [TD]Stoke[/TD] [TD=align: right]1.29[/TD] [TD]H[/TD] [TD][/TD] [TD]Standard Error[/TD] [TD=align: right]0.07[/TD] [TD][/TD] [TD]Standard Error[/TD] [TD=align: right]0.07[/TD] [TD][/TD] [/TR] [TR] [TD]E0[/TD] [TD=align: right]03/11/2012[/TD] [TD]Man United[/TD] [TD]Arsenal[/TD] [TD=align: right]1.69[/TD] [TD]H[/TD] [TD][/TD] [TD]Median[/TD] [TD=align: right]1.29[/TD] [TD][/TD] [TD]Median[/TD] [TD=align: right]1.27[/TD] [TD][/TD] [/TR] [TR] [TD]E0[/TD] [TD=align: right]24/11/2012[/TD] [TD]Man United[/TD] [TD]QPR[/TD] [TD=align: right]1.18[/TD] [TD]H[/TD] [TD][/TD] [TD]Mode[/TD] [TD=align: right]1.29[/TD] [TD][/TD] [TD]Mode[/TD] [TD=align: right]1.29[/TD] [TD][/TD] [/TR] [TR] [TD]E0[/TD] [TD=align: right]28/11/2012[/TD] [TD]Man United[/TD] [TD]West Ham[/TD] [TD=align: right]1.24[/TD] [TD]H[/TD] [TD][/TD] [TD]Standard Deviation[/TD] [TD=align: right]0.24[/TD] [TD][/TD] [TD]Standard Deviation[/TD] [TD=align: right]0.24[/TD] [TD][/TD] [/TR] [TR] [TD]E0[/TD] [TD=align: right]15/12/2012[/TD] [TD]Man United[/TD] [TD]Sunderland[/TD] [TD=align: right]1.2[/TD] [TD]H[/TD] [TD][/TD] [TD]Sample Variance[/TD] [TD=align: right]0.06[/TD] [TD][/TD] [TD]Sample Variance[/TD] [TD=align: right]0.06[/TD] [TD][/TD] [/TR] [TR] [TD]E0[/TD] [TD=align: right]26/12/2012[/TD] [TD]Man United[/TD] [TD]Newcastle[/TD] [TD=align: right]1.24[/TD] [TD]H[/TD] [TD][/TD] [TD]Kurtosis[/TD] [TD=align: right]-0.31[/TD] [TD][/TD] [TD]Kurtosis[/TD] [TD=align: right]0.25[/TD] [TD][/TD] [/TR] [TR] [TD]E0[/TD] [TD=align: right]29/12/2012[/TD] [TD]Man United[/TD] [TD]West Brom[/TD] [TD=align: right]1.29[/TD] [TD]H[/TD] [TD][/TD] [TD]Skewness[/TD] [TD=align: right]1.12[/TD] [TD][/TD] [TD]Skewness[/TD] [TD=align: right]1.36[/TD] [TD][/TABLE] If you look @ utd's home win odds they have a positive Skew (meaning the majority of odds are on the left with a positive tail to the right due to outliers) When this is the case (either positive or negative) the mode should be used as the central limit indicator.(basic example) Therefore with 95% confidence the correct odds at home for utd should lie within 1.14 to 1.42 (mode +/- 0.15) and not the mean 1.22 to 1.52 So in effect by thinking you have 'value' backing odds over 1.42 you are infact heading for the black swan effect. Nb Before anyone states the obvious that the skewness is within the accepted 2 ses (standard errors of skewness) of 1.41 for the number of variables, 1.36 is imo considerable enough to be taken into consideration. :ok