Probability to Odds rule on page 24 with m = 1). I found n by taking the reciprocal of the probability. For example, the probability of “3 white balls” is .003837 and its reciprocal is Thus the probability of “3 white balls” is approximatelyand the odds are 1:260. The odds for losing were determined by first calculating the odds of winning something. The probability of winning something is about 1 - 0.9723 = 0.0277, because 0.9723 is the probability of “None of the above.” Since he

term. Betting Strategies • If, in your judgment, the true odds t:100 for the favorite are better than the offered odds (u + s) : 100, then bet on the favorite. • If, in your judgment, the true odds t:100 for the favorite are worse than the offered odds (u - s) : 100, then bet on the underdog. • Otherwise, don’t bet. The “don’t bet” line reflects the fact that your belief is too close to that of the house. The little edge you might have is swamped by the service charge. Starting on

See page 131, where I provide some imaginary data. We would expect pitchers’ winning percentages and their earned-run-averages, ERAs, to have a negative correlation, because the better pitchers should have higher winning percentages and lower ERAs. For the top 50 pitchers in 2003, based on their ERAs, the correlation was negative, as expected, namely ≈-0.347, but this is a surprisingly weak correlation. The correlation of ERAs with total wins was even weaker, an almost insignificant ≈-0.160. I

simulations of the 1941 season, using the baseball computer game, DiamondMind Baseball. The five longest streaks in that 100 years of 1941 baseball were 55, 52, 40, 39, and 36, the last three of which belonged to DiMaggio. All told, DiMaggio had the longest hitting streak in 16 of the 100 seasons. In any case, these results suggest that DiMaggio’s 56-game hitting streak was remarkable, but that very long streaks are to be expected every few decades. It is interesting to note that the Yankees won

if the correlation is strong. 6 This is similar to the following remark. Because I don’t know your birthdate, each date has very small probability of being your birthdate (≈0.003), yet one of those dates must be your birthdate. 7 Trademarks of Competitive Edge Decision Systems. 8 Normalizing a statistic involves multiplying and/or adding constants to be able to compare the statistic with some standard statistic. 9 For example, the established level projected the 2005 values PTS within