Where version 6 gains or loses against the 2006 table: an information-only breakdown of the version 6 CHF blind test, with the Ruan diagnostic

An information-only, post hoc breakdown of the version 6 blind test of critical heat flux in uniformly heated round tubes (10.5281/zenodo.23994262). It asks where, inside the 370 scored primary points, version 6 does better or worse than the 2006 look-up table, and whether the large miss on Ruan's TU Berlin points is shared across methods. It decides nothing and changes no verdict of version 6, version 7 or the standing all-versions rule. On this set the conditions and the laboratories coincide, so it cannot tell a regime effect from a laboratory effect; the main results are within each article. The plan was approved by the author before any result of this analysis was computed or seen, and the analysis was run once. Within each article, the table has the lower relative RMS error in all but one group of Previti and in 7 of 9 groups of Wurtz; version 6 is lower in 8 of 9 groups of Ruan. On Ruan's 82 points every method's mean signed error is negative (table -22.5%, C -22.1%, version 6 -15.8%), and the errors of the three methods that rest on the table move together. Setting Ruan's heated length to 0.399 m instead of 0.383 m raises every method's relative RMS error, by 0.2 to 2.4 points. A reading of Ruan's method pages records that CHF was taken during a slow power ramp with a fixed wall-temperature cut-off. The author adopted six candidate hypotheses for a later blind set; none is a claim until frozen and scored once. The analysis code, every output, the approved plan and the method record are included, with the faults found by independent checks stated plainly. Nothing here supports any use in design, safety or licensing. Archive SHA-256: cc4fc3b5f404890807d1f7eaf0b07da8440cbf3e2d51248ad47e31a566914b3a Source documents were obtained through the Interlibrary Loan service of J.Y. Joyner Library, East Carolina University, by the author as an ECU student.