---
title: "Re-read GAO/IMTEC-92-26 to verify verbatim the Patriot Dhahran time-drift mechanism (24-bit / 0.1, 0.3433 s, 687 m) and the Feb 11–26 warning/patch timeline"
type: "question"
status: "answered"
date_raised: "2026-07-11T00:00:00.000Z"
writer_model: "claude-opus-4-8"
tags: ["verification","patriot-missile","gao-report","rounding-error","gulf-war"]
answered_log: "2026-08-26 — answered by 10-inbox/raw/2026-08-26-re-read-gaoimtec-92-26-to-verify-verbatim.md, an independent re-fetch of the gao.gov primary (via a Wayback Machine capture; gao.gov 403s tooling directly) that confirmed all five items verbatim. (1) Mechanism quote and footnote 4 confirmed — audit_status of [[claim-patriot-dhahran-1991-clock-drift-from-24-bit-truncation-of-tenths]] updated. (2) Appendix II's full six-row progression, including the exact 100-hour figures and units, confirmed and recorded in new note [[claim-patriot-dhahran-1991-time-drift-proportional-to-runtime-appendix-ii-table]]. (3) Feb 11 warning quote confirmed — audit_status of [[claim-patriot-dhahran-1991-known-bug-patch-arrived-one-day-late]] updated; the underlying Israeli-data-recorder provenance is now its own note, [[claim-patriot-dhahran-1991-israeli-data-recorders-diagnosed-drift]]. (4) Feb 26 patch-arrival quote confirmed, with a fuller, previously-elided clause recovered ('from the United States to all Patriot locations') — same note's audit_status. Two previously uncaptured intermediate dates (Feb 16 software release, Feb 21 general warning) also surfaced and are recorded in [[claim-patriot-dhahran-1991-feb16-fix-feb21-warning-lacked-runtime-guidance]]. (5) The 28-casualty figure confirmed verbatim in both the primary's transmittal-letter and body phrasings. What settled it: a second, independent fetch of the same document via a different route (Wayback Machine vs. the original cs.nyu.edu mirror), landing on identical text throughout."
---


Two claim-notes promoted 2026-07-11 rest entirely on a single Tier 1 primary — the U.S. General Accounting Office report *GAO/IMTEC-92-26* — that the headless promotion run could not independently re-fetch (no web tool). Both are marked `audit_status: capture-verified` pending an independent verbatim re-read. This question routes that re-check into the nightly queue.

- [[claim-patriot-dhahran-1991-clock-drift-from-24-bit-truncation-of-tenths]]
- [[claim-patriot-dhahran-1991-known-bug-patch-arrived-one-day-late]]

## What to pull

Fetch the report (freely mirrored at https://cs.nyu.edu/~exact/resource/mirror/patriot.htm; cross-check against the GAO's own archive of IMTEC-92-26) and confirm, verbatim:

1. **Mechanism quote** — "Because of the way the Patriot computer performs its calculations and the fact that its registers are only 24 bits long, the conversion of time from an integer to a real number cannot be any more precise than 24 bits."
2. **Appendix II data row** — that at 100 hours of continuous operation the time error is **0.3433 seconds** and the range-gate shift is **~687 meters**. Confirm the exact figures and units, and whether 687 m is stated or derived.
3. **Warning quote** — "On February 11, 1991, the Patriot Project Office received Israeli data identifying a 20 percent shift in the Patriot system's radar range gate after the system had been running for 8 consecutive hours."
4. **Patch-arrival quote** — that the "modified software ... arrived in Dhahran" on **February 26**, one day after the 25 February strike, and the logistics attribution: "the delay in distributing the software ... was due to the time it took to arrange for air and ground transportation in a wartime environment."
5. **Casualty figure** — that the report (or a corroborating primary) states **28** killed at the Dhahran barracks.

## Why it matters

The quantitative claims (0.3433 s, 687 m, 20%, 8 hrs, 24 bits) and the one-day timeline are the load-bearing content of both notes; the sourcing floor puts numbers and mechanism at Tier 1–2, which the GAO report clears. What is missing is only the independent re-fetch. Resolving this promotes both notes from `capture-verified` toward `verified-verbatim` and clears them to move past `seedling`.


## Progress log

- 2026-08-26 — answered by 10-inbox/raw/2026-08-26-re-read-gaoimtec-92-26-to-verify-verbatim.md, an independent re-fetch of the gao.gov primary (via a Wayback Machine capture; gao.gov 403s tooling directly) that confirmed all five items verbatim. (1) Mechanism quote and footnote 4 confirmed — audit_status of [[claim-patriot-dhahran-1991-clock-drift-from-24-bit-truncation-of-tenths]] updated. (2) Appendix II's full six-row progression, including the exact 100-hour figures and units, confirmed and recorded in new note [[claim-patriot-dhahran-1991-time-drift-proportional-to-runtime-appendix-ii-table]]. (3) Feb 11 warning quote confirmed — audit_status of [[claim-patriot-dhahran-1991-known-bug-patch-arrived-one-day-late]] updated; the underlying Israeli-data-recorder provenance is now its own note, [[claim-patriot-dhahran-1991-israeli-data-recorders-diagnosed-drift]]. (4) Feb 26 patch-arrival quote confirmed, with a fuller, previously-elided clause recovered ('from the United States to all Patriot locations') — same note's audit_status. Two previously uncaptured intermediate dates (Feb 16 software release, Feb 21 general warning) also surfaced and are recorded in [[claim-patriot-dhahran-1991-feb16-fix-feb21-warning-lacked-runtime-guidance]]. (5) The 28-casualty figure confirmed verbatim in both the primary's transmittal-letter and body phrasings. What settled it: a second, independent fetch of the same document via a different route (Wayback Machine vs. the original cs.nyu.edu mirror), landing on identical text throughout.