Tasmania, Australia

Christopher Beach

I've been building software for around thirty years.

I run Books.org, a community for readers. I sit on the board of the International Space Association, where I've served since 2018, and I'm Strategic Technical Advisor to Work Metrics.

I'm interested in astronomy and mathematics. On the astronomy side I built AstroSat, a satellite trail forecaster for astrophotographers. In mathematics, my earlier interests included Markov chains, an early precursor to AI large language models, and I recently published a proposed proof of Kotěšovec's conjecture about OEIS sequence A188115.

Portrait of Christopher Beach

Proposed proof, recently published

A proposed proof of Kotěšovec's conjecture for OEIS sequence A188115

Take the whole numbers from −n to n and leave out zero. Now choose n different ones that add up to exactly zero. When n is 3 there are two ways: −3, 1, 2 and −2, −1, 3. Sequence A188115 in the On-Line Encyclopedia of Integer Sequences counts the ways for each n, and the counts climb quickly. From n = 2 they run 2, 2, 8, 16, 52, 152, 484.

The exact counts take real work to compute, so it helps to know roughly how big they get. Václav Kotěšovec conjectured that for large n the count is close to

√3 × 4n / (π × n2)

which means each step up in n eventually multiplies the count by a little under four. The formula tracked the computed values, but that is not the same as knowing it always will. My proposed proof shows that it does: the larger n gets, the smaller the percentage error, and it shrinks towards zero.