Fresh frozen stellar hierarchy transfer exposes missing-input recall limits

Does the frozen hierarchical detector retain sufficient precision and recall on new unfiltered survey directions?

Four new WISE All-Sky rectangles and independent SDSS primary spectra; unchanged E124/E112 models, mappings and thresholds, raw baselines and no test-selected winner. Preserve all requested targets, unknowns, control alternatives, area burden and review budgets.

Eight bounded GETs, independent DOM/Decimal 345332 native primitives and69227 null states. Independent60960 features,45720 method/source records,214486 inference scalars (max1.066e-13),1155992 geometry decisions/266 alternatives,304 audits,30 budgets and345 curves.1176 native review primitives.

7620 new native sources/304 spectra, zero development overlaps; response11.25MB. Primary7/14 stars+2/61 others (77.78%precision/50%recall), secondary8+5 (61.54%/57.14%), E1127+3. All finite gates fail, with only14 positive proxies.2186/2908/2230 candidate entries retained. Six unknown proxies are stars with W2SNR<5; field requested21 stars comprise7 no alternative,6 unscorable,1 primary rejection,7 selected. Support ceiling8/21; catalogue ceiling14/21. No retuning.

Next: Keep this failed transfer result and all original star/YSO/dust/103 molecular tiers. Broaden compatible optical and signed low-SNR IR views with uncertainty-aware label-free embeddings, preserve confusion/requested denominators and freeze before a distinct survey cohort. Dust spatial gates/embedded contaminants remain unresolved; E121 source-list upload still pending explicit permission.

Sparse target-selected spectra cover0.984%of sources; wide Wilson intervals and deliberately fixed regions do not establish all-sky purity. Optical/IR selection functions differ; static proximity omits epoch/blend ownership. Supported recovery7/8 must not replace7/14 or7/21. This cohort becomes development evidence after assessment. No physical class/birth or adopted replacement.