It is 1918, and a geneticist has crossed plants and counted the offspring types — but the cross yielded only a few dozen progeny, and from those counts you must estimate how tightly two genes are linked on the chromosome. The estimate is a maximum-likelihood exercise on small integer counts, and the sampling variance at this size is large enough to swallow a real effect or manufacture a false one. Deliver the estimate and its precision, and say honestly whether the two genes are linked at all. Get it wrong and a genetic map is drawn with phantom order, or true linkage is dismissed — and at the dawn of quantitative genetics the map itself is the science.
Hopper's pioneering work on the A-0 compiler at Remington Rand in the early 1950s and her leadership in developing COBOL made machine-independent programming languages possible, a foundational contribution to computer science but one entirely disconnected from statistical estimation or genetics. She built tools for translating human-readable instructions into machine code, not for fitting likelihood functions to biological count data or reasoning about the precision of a small-sample estimator. Her career at the US Navy and in industry spanned computing infrastructure and software engineering, decades removed from and methodologically unrelated to the 1918 quantitative-genetics setting of this problem. There is no meaningful bridge between compiler design and maximum-likelihood estimation of a recombination fraction; even framed generously, Hopper's actual expertise offers essentially nothing toward solving a small-sample genetics inference problem, however historically important her broader contributions to computing remain.
Wainwright, at Berkeley since the 2000s, works on high-dimensional statistics, graphical models, and variational inference, with a research program explicitly focused on characterizing how estimation difficulty scales with dimension and sample size — a genuinely relevant theoretical lens for understanding why a tiny sample makes the recombination-fraction estimate unreliable. His work on minimax rates and concentration inequalities could in principle be used to derive sharp, non-asymptotic bounds on the estimator's precision rather than relying on classical asymptotic Fisher-information approximations that are shaky for a few dozen counts. But his actual body of published research targets modern high-dimensional and graphical-model problems, not classical low-dimensional genetics estimation, and he has no direct engagement with Mendelian linkage analysis. The theoretical instincts transfer better than the specific tools, making him a moderate rather than strong match for this exact historical problem.
Battle #131 · 8/10/2026, 11:39:20 AM · this result is deterministic: the same two personas on this problem always resolve the same way.