Add cohort record and resolution figure

README.md

@@ -50,6 +50,10 @@ Reconstruction is stochastic and imperfect. A manual audit by a single annotator
 
 Every reported score therefore carries an interval, and both come from the same Monte Carlo run: each draw perturbs the released labels at the audited rates and scores the metric against them. The point value is the mean over the draws, the interval the spread between their 2.5th and 97.5th percentiles. A wide interval means a comparison does not survive the label error, not that the measurement failed.
 
+The width of that interval is the benchmark's resolution κ, the smallest difference in true performance it can distinguish under its label error. A reported gap below κ does not separate two detectors. Because the interval widens with detector skill, κ depends on where on the quality scale it is read: at node granularity it stays below 0.04 for F1 up to an AUC of 0.9 and reaches 0.12 in the worst case.
+
+![Resolution across detector skill at node granularity](resolution.svg)
+
 ## Protocol
 
 For each case a submission holds zero or more locations on changed lines, each given as a path, a line number, and the side of the diff it refers to. An empty submission states that no defective line was found, and a case left out entirely is treated the same way.