Monotonicity conjecture for favorable longer-word pairs

Let nn be a sufficiently large positive integer. For binary words v,wv,w of length at most nn, let τv\tau_v and τw\tau_w be their first occurrence times in an iid Bernoulli(p)(p) sequence, and let vv be longer than ww.

Monotonicity conjecture. The number of pairs (v,w)(v,w) satisfying

Pp(τv<τw)>12\mathbb{P}_p(\tau_v < \tau_w)>\frac{1}{2}

is non-increasing as a function of p(0,12)p\in\left(0,\frac{1}{2}\right).

This conjecture formalizes the expected increase, as the Bernoulli parameter decreases, in the frequency with which a longer word defeats a shorter word. It is supported by computer computations but is disproved by the example (v,w)=(100010,001100)(v,w)=(100010,001100), whose win probability is 12(1p(1p)4)\frac{1}{2}(1-p(1-p)^4) and is minimized at p=15p=\frac{1}{5}, so the asserted monotonicity fails.

Sources & referencesView supporting material

Primary source

Mathew Drexel, Xuanshan Peng and Jacob Richey, “Word length, bias and bijections in Penney's ante”, arXiv:2409.19195 (2024).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.