from Hacker News

The Statistics of Coin Tosses for Theater Geeks

by aet on 4/27/17, 11:27 AM with 5 comments

  • by Zanni on 4/27/17, 10:28 PM

    The article shies away from the fascinating question it hints: given the improbability of 92 consecutive tosses landing on heads, what are the odds that something else is at work here? (un-, sub- or super-natural forces, as they put it in the play). That is, at what point is the improbability so absurd that something else is more likely to be true? That the coin is biased. That Rosencrantz is lying. That laws of probability are actually not in effect. As it happens, this last is true. They're not. R&G are in a play, and the spin of the coin is controlled by Tom Stoppard, the playwright, and not probability. So how improbable does something have to become before you suspect that you're in, e.g., a simulated universe?
  • by anateus on 4/27/17, 7:04 PM

    Some may be interested in Von Neumann's algorithm for getting fair results even from a biased coin: Toss it twice. If the results are the same, ignore. If they are different, use the first coin's result.