Exact Depth-Two and Large-Depth Guarantees for Recursive Rejection Sampling
Year: 2026
Type: preprint
Abstract: Proves the optimized uniform/dead binary witness is strictly suboptimal for every fixed draft count at sufficiently large depth. A subtree deletion improves it by an explicitly bounded exponential amount.
Author: Anonymous
Language: English
Subfield: Computational Theory and Mathematics
Field: Computer Science
Domain: Physical Sciences
Keyword: exponential bounds
Open Access status: green
OpenAlex ID: W7236235143
