# Desk log CUPED-PRE-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 51,800-account pack split 50%/50%. Prior 28-day orders had 6% null coverage; the unadjusted and CUPED deltas were 0.9 and 0.8 points, with refunds labeled unchanged. CUPED for new users and the mixed-population decision remained held.

The [CSV](https://infinisynapse.com/blog-media/cuped-explained/downloads/covariate-summary-CUPED-PRE-20260823.csv) exposes seven aggregates and two held items. The [verification script](https://infinisynapse.com/blog-media/cuped-explained/downloads/verify-CUPED-PRE-20260823.py) checks displayed output only. Rows, timestamps, covariance, correlation, theta, variances, intervals, code, and estimands are unavailable.

External context: [Deng et al., Improving the Sensitivity of Online Controlled Experiments by Utilizing Pre-Experiment Data](https://www.microsoft.com/en-us/research/publication/improving-the-sensitivity-of-online-controlled-experiments-by-utilizing-pre-experiment-data/), [ASA p-value statement](https://www.amstat.org/asa/files/pdfs/p-valuestatement.pdf), and [NIST/SEMATECH e-Handbook](https://www.itl.nist.gov/div898/handbook/) (retrieved 2026-08-31). These publishers did not review this run.