Hi HN,

I just hit a major milestone on my side project.

I work a demanding day job, so I built a platform where Stockfish plays continuous self-play (2s/move to keep server load manageable) to analyze asymmetrical chess variants.

In Chess960v2, White and Black independently choose any of the 960 valid setups, resulting in 921,600 possible starting combinations. I'm running a massive round-robin tournament (960 rounds total, where each setup plays against every other setup), and 600 out of 960 rounds have been completed.

The goal is to mathematically find which combinations are genuinely balanced (50/50) and which give an unfair engine bias to one side due to asymmetry.

I'm pretty exhausted today, so I might be slow to reply, but I'd love for you to check out the site, look at the current data, and share your thoughts.

Thanks!