History
Success probability of optimal nested scalable oversight in the Backdoor Code game at a 400-point general Elo gap
Reassessed
First pass after the split. Kept the existing definition, which is already complete. Recorded the 10.0% figure from the Engels et al. 2025 abstract. Estimated a central value of 0.10 with a 90% interval of 0.05 to 0.20, widened by judgment for fit and extrapolation uncertainty. Set importance to 0.3. The pass was limited by budget, so the full PDF was not read.
Claim entered the graph