Confirmation of theory

There is evidence around that you can earn money from correctly forecasting a certain amount of probability. You can even buy shares listed on the London stock exchange in companies that do this and return value to develop there business and the returns to their shareholders.

There is also other compelling evidence that with careful and considered staking in pari-mutuel systems you can overhaul, with knowledge, the odds of winning a reasonable amount of money. By forming a syndicate, you can bring closer that one in a lifetime chance of winning a big amount on a betting system. While the much longer term odds of getting a good margin are poor, the chance of winning "A big one" or winning consistent amounts within a finite time scale are possible. The following extract from The Economic Journal July 1996 (Compelling reading for me but not many other people) confirms the inconsistencies with pari-mutuel systems.

Of the lengthy report entitled, Optimal betting and efficiency the closing chapter is such:-

EXTRACT of The Economic Journal - July 1996

III. Conclusion

The description of the market and data from the Woodlands horse and dog races suggests a market in which information costs lead to differences in information across bettors. Section II of this paper shows that the presence of information costs combined with takeouts by tracks in pari-mutuel games leads to expected returns and market odds differing substantially from the perfect information betting market.

Furthermore, the model provides predictions consistent with findings of previous empirical studies. The wagers of informed bettors push odds higher on bets with low expected values and explain the low expected values and explain the low expected return to such bets found in empirical studies. Optimal purchases of information and wagers lead to market odds which fail to reflect true probabilities of outcomes and expected returns which may exceed one dollar on a one dollar bet in some races. Expected returns converge to one for track takeouts near zero and small information costs relative to the size of total pool, as in most financial markets. This result suggests that pari-mutuel games with high takeouts provide poor approximations of these markets. The model suggests that, unlike investors in most financial markets, optimal purchases of information by agents at racetracks leads to market odds which fail to directly reflect the true probabilities of outcomes in races.