Coin-flip odds – Letter to the Editor
We have to agree with most conclusions in Phil Mole’s piece “Are Skeptics Cynical?” (November/December 2002). However, Phil Mole is no statistician, and his statistical examples on pages 45 and 46 are incorrect. The statistical method that has to be applied in such cases is the “binomial formula.” Indeed, it confirms our commonsensical expectation that if we flip fair coins, then an outcome close to 50 percent head and 50 percent tail is much more probable than numerous heads or tails in one row. The exact probabilities can be easily calculated from the binomial formula, or can be looked up in binomial tables found in many elementary statistical textbooks. The binomial formula is somewhat similar to that used by Phil Mole, but with the inclusion of the so-called binomial coefficient. In Phil Mole’s examples these probabilities are:(l) 0.2461, (2) 0.0098, a 25-times difference, and (3) 0.2734, (4) 0.0312, an almost 9-times difference in probabilities. Those who chose (1) and (3) as being the most statistically likely are really in good company.
George K. Nagy
Albany, New York
COPYRIGHT 2003 Committee for the Scientific Investigation of Claims of the Paranormal
COPYRIGHT 2003 Gale Group
