Quote:
|
However, if the event is performed enough times in succession the law of averages tells you that the past results do in fact influence future outcomes in that past annomolies in one direction will correct over time by compensating in the other direction and returning to the average.
|
For a 50/50 coin flip, you're looking at ~100 flips to have a good sample size. So for this person's 40/60 "coin flip" over 6 "flips" it's telling a very unreliable story.
Because this is a typically "one and done" tailoring combine, you're not going to see that leveling out over a large sample size (example: having 6 combine
successes in a row). So people's perceptions are skewed.
Basically the lower your attempt / sample size, the less power the "law of averages" will affect you. It's much much more likely to have skewed results with 4 coin flips than with 1,000 coin flips.