4
$\begingroup$

In Time Travel and Other Mathematical Bewilderments, Martin Gardner presents a set of four nontransitive bingo cards designed by Donald Knuth (pp. 61). The rules are that the first player to complete a horizontal row wins. Gardner does not delve into the mathematics but merely mentions that, probabilistically, A beats B, B beats C, C beats D, and D beats A.

enter image description here

I was so baffled that I immediately went on to to verify those results. He was right.

Now, with dice or heads tails sequences, I can understand nontransitivity; can someone please give the mathematical explanation of how and why nontransitivity occurs in the above game.

$\endgroup$
  • 1
    $\begingroup$ An observation: Where card $x$ beats card $y$, the numbers on card $x$ that are not on card $y$ are in different rows, but the numbers on card $y$ that are not on card $x$ are in the same row. This gives card $x$ the advantage, although my proof is too long to fit in this comment. $\endgroup$ – Steve Kass Feb 27 '15 at 21:55
  • 1
    $\begingroup$ @SteveKass You and Fermat would have gotten along. :) $\endgroup$ – eigenchris Feb 27 '15 at 22:27
  • $\begingroup$ What does it mean for one card to beat another one? Is there a payoff involved? Because clearly there are some situations where B will win before A, etc. $\endgroup$ – Ibrahim Tencer Feb 27 '15 at 22:59
  • 2
    $\begingroup$ @IbrahimTencer, the numbers $1$ through $6$ are called out in random permutation. One card "beats" another if it wins more of the permutations. $\endgroup$ – Barry Cipra Feb 27 '15 at 23:05
  • $\begingroup$ I changed the wording; thanks for that. $\endgroup$ – blackened Feb 27 '15 at 23:06
1
$\begingroup$

Let's say two players take cards $A$ and $B$ and we roll a 6-sided die to call out the numbers.

If a $1$ or a $3$ is rolled right off the start, then player $B$ cannot win using their top row and must rely on a $5,6$ victory. Player $A$ does not experience this problem if something on player $B$'s card is rolled first.

$\endgroup$
  • $\begingroup$ Nice observation. That should be the explanation. $\endgroup$ – blackened Feb 27 '15 at 23:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.