'Gambling Shock' baseball system

[Orpheus]

The basis of this system is that there are two types of shock in gambling (see William Mallios' book The Analysis of Sports Forecasting: Modelling Parallels between Sports Gambling and Financial Markets for further details):

• Statistical shocks, which are differences between the result of a match and its theoretical prediction from a true statistical model
• Gambling shocks, which are differences between the result of a match and its expected outcome given the odds on the match.

Statistical shocks are difficult to gauge as the derivation of a 'true statistical model' of very difficult, if not impossible. However, it is quite easy to see how a gambling shock would have an affect on future team performance. For example, Liverpool lost 2-1 at home to Barnsley in the FA Cup last year; they then won their next seven matches, including home and away to Inter Milan in the Champions League. Conversely, Barnsley failed to win again in their next four matches. So, the reasoning is thus: Short-priced favourites who lose (home or away) will not lose the next time that they are in the same situation at such a short-price; Large-priced underdogs who win (home or away) will not win the next time that they are in the same situation. This is a one-lagged effect of a gambling effect. For example, in baseball this year, home favourites who are 1.5 or lower and lose are 11-0 the next time that they are home favourites and 1.5 or lower. The benchmark odds are the following: Home favourite [HF1]: 1.5 Road favourite [RF1]: 1.8 Home underdog [HD1]: 2.2 Road underdog [RD1]: 2.8 These parameters have been derived using historic odds data within a one-step-ahead forecasting model (as opposed to finding parameters that give the best result on historic data, i.e. 'data-fitting') There may be circumstances in which a reaction to a gambling shock may not lead to a positive result in the next equivalent match (performance is not always equated to result), plus there is a myriad of other factors that cause gambling odds to not reflect the true probabilities of a match result (e.g. bookies look to balance their liabilities on a game rather than better predict the result of a match than the typical punter) and this is the cause of the difference between gambling shocks and statistical shocks. Despite home advantage in football, teams do not continually win their home matches against weaker teams, etc. So, I separately model two-, three-, four- and five-lagged effects of gambling shocks. In these cases, the correction to the shock will be made, but it may take longer to materialise in terms of the result. The problem is that beyond the game immediately after shock result, it is difficult to disentangle the various reasons why 'corrections' may occur. So, these are modelled as lagged gambling shocks, but an explanation of the reasoning behind them is harder to make. The benchmark odds for the longer-lagged gambling shocks are the following: Home favourite [HF2-5]: 2.2 Road favourite [RF2-5]: 1.8 Home underdog [HD2-5]: 2.2 Road underdog [RD2-5]: 1.6 The deriation of these parameters is as above. Note the 'perverse' benchmark odds for HF2-5 and RD2-5. If a home team has been 2.2 or better and lost at least twice in a row at those odds, I back them when they are 2.2 or better at home again. This may be a home advantage effect that is underestimated by the odds compiler, for example ... weak teams will win at home. For RD2-5, this will be converse case, favourites on the road will lose because of home advantage, for example, and their winning streak on the road (which led to lower odds) will end. If gambling shocks have any lagged effects, this system should produce a profit. I will outline this for football at the start of next season, but I will concentrate on baseball for now. To recap, if the odds are lower than the benchmark odds for the home favourite or road favourite team (and they have lost their last 1-5 times in similar circumstances) they are backed. If the odds are higher than the benchmark odds for the home underdog or road underdog teams (and they have won their last 1-5 times in similar circumstances) they are opposed.

[Orpheus]

Re: 'Gambling Shock' baseball system Tonight's games: RF1: Kansas City HF2: Washington RF2: Chicago Cubs RD2: against LA Angels; against San Francisco HF3: Chicago White Sox; San Diego HF4: Detroit RD5: against Philadelphia