Analytical model claims cannot exceed their calibration domain

A sensor or chemometric model should not be used to make a whiskey age, authenticity, classification, or quality claim beyond the products, sample states, procedures, and uncertainty represented in its calibration and validation data.

Model performance remains bounded by sample matrix, process states, target variables, reference methods, calibration range, and independent validation. SRC-111 adds a direct working-distillery example with separate rye-mash, barley-mash, and molasses models; it does not validate whiskey-age, authenticity, or sensory-quality claims.

Independent beer and fortified-wine chapters show brand- or product-specific signals, sample-state confounders, repeated-calibration uncertainty, preprocessing choices, and dependence on known reference sets.

SRC-107 supplied independent cross-category methodological evidence. Review of the supplied SRC-111 copy added explicit NIR calibration requirements and direct rye/barley-mash distillery applications; the bounded claim is now ready for drafting. September28 fresh audit read all448 supplied sheets, but printed22 is missing and equations332–333 have glyph limitations; full published-work completion is not claimed. The located qualitative evidence remains usable with its recorded qualifications. SRC-109 fresh review now covers1185/1185 sheets and443 detail inspections, including all appendices, glossary and index. Claim maturity and sample-specific limits are unchanged; completion is coverage evidence, not experimental replication. September 28–29 fresh review of Alcoholic Beverages covers all 552 sheets and 121 detailed visual pages. SRC-107 and SRC-259 are catalog records of the same edition, not independent corroboration. Existing evidence was rechecked with chapter-specific error/transfer limits; claim confidence and other-source review status are unchanged.