Ok, this isn't about blackjack, but I see some math correlations here, so bear with me. Many of you may be familiar with a study being done on estrogen supplements such as Prempro and Premarin for non-menopausal reasons,(estrogen as vitamin, if you will). My girlfreind happens to work for the company that makes Prempro and her company is taking a good PR hit for this,(although she says this kind of thing is common.) Anyways, the numbers looked scary. Heart attack rate; 29 percent higher among the Prempro group than the control group. Stroke rate was 41 percent higher, and risk of invasive breast cancer was 26 percent higher. However, these numbers are translated from the fact that out of 10,000 healthy women from 50 to 79, 37 women were struck with heart attacks versus 30 for the control group. Strokes were 29 vs. 21, and breast cancer was 38 vs. 30(Numbers from Newsweek, 7/22/2002). These numbers come from an 8-year study of 16,608 women which was cut off after 5 years. Now knowing that 100,000 hands of blackjack is nowhere near long run, can these numbers really be statistically significant in comparison? I suppose if blackjacks went up 26% over 100,000 hands, that it would denote something amiss, but what is the actual criteria for significance here? At what point is the threshhold crossed? Five percent higher? I know SW had something about this in his SSB book, although obviously for different circumstances, but I figured I'd go straight to the pros about this. So, SW, MathProf, and anyone else who'd like to tackle this, please straighten me out. By the way, my 2 cents on this is; these numbers are all lower than the same inferences for obesity, smoking, and drinking, but not too many people who have those conditions are concerned.
