# Desk log ABT-GUARD-20260823

**Status:** First-party illustrative assignment composite; not customer, production, randomized-trial, benchmark, or independently audited data  
**Run date:** 2026-08-23  
**Evidence checked:** 2026-08-31  
**Operator:** InfiniSynapse Data Team  
**Attestor:** William Zhu, InfiniSynapse cofounder ([GitHub @allwefantasy](https://github.com/allwefantasy))

The operator inspected an illustrative 42-day table: 86,000 units split 50.1%/49.9%, a 1.8-point primary delta, and a 0.4-point refund guardrail delta. Prior 28-day orders were declared as a CUPED covariate but are unavailable. The ship decision remained held.

The [CSV](https://infinisynapse.com/blog-media/ab-test-analysis/downloads/experiment-summary-ABT-GUARD-20260823.csv) exposes five aggregates and two held items. The [verification script](https://infinisynapse.com/blog-media/ab-test-analysis/downloads/verify-ABT-GUARD-20260823.py) checks displayed output only. Assignment/outcome rows, denominators, variances, confidence intervals, code, stopping rule, SRM test, and covariate diagnostics are unavailable.

External context: [ASA Statement on Statistical Significance and P-Values](https://www.amstat.org/asa/files/pdfs/p-valuestatement.pdf), [NIST/SEMATECH e-Handbook](https://www.itl.nist.gov/div898/handbook/), and [FTC Advertising Substantiation](https://www.ftc.gov/legal-library/browse/advertising-marketing/advertising-substantiation) (retrieved 2026-08-31). These publishers did not review this run.