That Random Coin Toss? Not So Random Afterall...

from the a-weaker-man-might-be-moved-to-reexamine-his-faith dept

One of my all-time favorite scenes in a play and movie, is the scene in Tom Stoppard's Rosencrantz & Guildenstern Are Dead where every coin toss comes up heads, leading to a bit of a philosophical discussion on probability. Of course, the randomness of the coin toss is the quintessential example of a random event and is used regularly for a variety of situations in which randomness is required, let alone expected. Except... it turns out the common wisdom may be wrong. Paul Kedrosky has the news of a test that showed that if you ask people to try flip a coin and get more heads than tails, they will, and not by a small margin either. In the test, 13 people were asked to flip a coin 300 times, trying to get as many heads as possible. All 13 participants got more heads than tails. Seven out of the thirteen had statistically significant margins of heads over tails (meaning almost certainly not a matter of chance). The highest was one individual had 68% of the coin flips land heads. In other words, a coin toss isn't particularly random.

Filed Under: coin toss, random


Reader Comments

Subscribe: RSS

View by: Time | Thread


  1. identicon
    Anonymous Coward, 13 Dec 2009 @ 11:12am

    Measuring constants to determine probability still moves the goal post only 1 to 2%

    Section 19.2 describes how to cheat at a coin-toss:

    http://bayes.wustl.edu/etj/science.and.engineering/lect.19.pdf

    There exist coin-toss programs that remains slightly more reliable because they are based on Numismatics which take into consideration variability of a coin's size, distribution of weight, all are factors which can affect the outcome, and are incorporated into the algorithm.

    You could get real crazy and incorporate environmental factors such as atmospheric pressure, wind speed and turbulence to further complicate your statistical prediction model, but at this point, you're measuring environmental constants which affect a 1 to 2% variability from say, a 49-51 split. Visual representations of these complex equations can be quite beautiful. Consider the Mandelbrot set.

    But why go to all this effort? Besides, you're measuring the probability, it is just that- probability, and not the event's outcome. The full accurateness of such a complex probability equation is truly non-computable until the actual event occurs. But attempting to go to such lengths to measure probability could at one point or another may influence the outcome, and that just ruins the fun.

Add Your Comment

Have a Techdirt Account? Sign in now. Want one? Register here



Subscribe to the Techdirt Daily newsletter




Comment Options:

  • Use markdown. Use plain text.
  • Remember name/email/url (set a cookie)

Follow Techdirt
Special Affiliate Offer

Advertisement
Report this ad  |  Hide Techdirt ads
Essential Reading
Techdirt Deals
Report this ad  |  Hide Techdirt ads
Techdirt Insider Chat
Advertisement
Report this ad  |  Hide Techdirt ads
Recent Stories
Advertisement
Report this ad  |  Hide Techdirt ads

Close

Email This

This feature is only available to registered users. Register or sign in to use it.