Do Lewis & Postol (1993) state verbatim that the videos show no unambiguous evidence Patriot destroyed even one Scud warhead, and that the Army's claim fell to 52% with only 40% 'highest confidence'?
This is the direct follow-up to question-verify-lewis-postol-1993-patriot-video-evidence-quotes, which flagged that two existing claim-notes — claim-lewis-postol-1993-no-video-evidence-of-patriot-scud-warhead-kill and claim-patriot-army-effectiveness-claim-collapsed-from-96-to-52-percent — carried audit_status: capture-verified from a 2026-07-09 hop session that read the paper directly but was never independently re-fetched, because that promotion ran headless with no web tool.
This session re-fetched the primary PDF directly (extract_pdf on scienceandglobalsecurity.org/archive/sgs04lewis.pdf, sha256:51b2ffee..., tls: verified, 63 pages) and read the extracted text at /Users/seek/seek/cache/sources/51b2ffeee1f0c11ce5997a24eb9aefccb3957da265efac4b1a5413678bba4a48.txt. Every quote below was run through quote_check against that extraction before being recorded.
A note on the extraction itself: pdftotext on this file drops inter-word spaces unpredictably in justified columns (e.g. "videoswe havebeenable", "45 out of47") and preserves end-of-line hyphenation (e.g. "40 per-\ncent") rather than rejoining split words. quote_check failed against the clean, normally-spaced reading of each quote and only returned grounded: true once the exact character stream — missing spaces and line-break hyphens included — was reproduced. This is a known artifact of the extraction tool, not a sign the wording differs from the document; the clean quotes below are what the phrase reads as when the PDF is actually viewed/printed, and each is followed by the literal grounded string quote_check confirmed.
Claim: Lewis and Postol (1993) state verbatim that the news-media videos they obtained contain no unambiguous evidence Patriot destroyed even one Scud warhead
The paper's opening abstract states: "The videos we have been able to obtain contain no unambiguous evidence that Patriot destroyed even one Scud warhead. Rather, they contain substantial evidence that Patriot's success rate was very low, and provide a record of Patriot performance that clearly is inconsistent with the US Army's current claim that Patriot destroyed 52 percent of the Scud warheads that it engaged."
"The videoswe havebeenable to obtain contain no unambiguous evidence that Patriot destroyed even one Scud warhead." — exact string confirmed
grounded: trueviaquote_checkagainst the rawpdftotextextraction (missing space between "videos" / "we" / "have" / "been" / "able" reproduces a pdftotext justification artifact, not a wording difference).
source_quote: "The videoswe havebeenable to obtain contain no unambiguous evidence that Patriot destroyed even one Scud warhead." source_url: "https://scienceandglobalsecurity.org/archive/sgs04lewis.pdf" source_sha: "51b2ffeee1f0c11ce5997a24eb9aefccb3957da265efac4b1a5413678bba4a48" source_title: "Video Evidence on the Effectiveness of Patriot during the 1991 Gulf War" source_author: "George N. Lewis and Theodore A. Postol" source_venue: "Science & Global Security, 1993, Volume 4, pp. 1–63" source_date: 1993 source_tier: 1
Historical/technical claim about what a primary source states, Tier 1 (peer-reviewed journal, authors' own analysis of evidence they gathered), quote confirmed grounded by direct extract_pdf read plus quote_check. This confirms and upgrades claim-lewis-postol-1993-no-video-evidence-of-patriot-scud-warhead-kill from capture-verified to independently re-verified — the wording it already carries is accurate to the primary text.
Claim: Lewis and Postol (1993) state verbatim that the Army claimed Patriot destroyed 52 percent of engaged Scud warheads, with only 40 percent of the claimed kills in the "highest confidence" category
The paper's introduction (p. 2) states in a single sentence, distinct from the abstract's restatement of the 52% figure: "The Army also claims Patriot destroyed 52 percent of the Scud warheads it engaged, but only 40 percent of the claimed warhead kills are in the 'highest confidence' category." (Footnote markers 2 and 3 in the original attach to "warhead kills" and "category" respectively, citing the Army's own briefing data as "classified.")
"Patriot destroyed 52 percent of the Scud warheads it engaged, but only 40 per-\ncent of the claimed warhead kills2 are in the "highest confidence" category.3" — exact string (line-break hyphenation and footnote markers preserved) confirmed
grounded: trueviaquote_check.
source_quote: "The Army also claims Patriot destroyed 52 percent of the Scud warheads it engaged, but only 40 percent of the claimed warhead kills are in the "highest confidence" category." source_url: "https://scienceandglobalsecurity.org/archive/sgs04lewis.pdf" source_sha: "51b2ffeee1f0c11ce5997a24eb9aefccb3957da265efac4b1a5413678bba4a48" source_title: "Video Evidence on the Effectiveness of Patriot during the 1991 Gulf War" source_author: "George N. Lewis and Theodore A. Postol" source_venue: "Science & Global Security, 1993, Volume 4, pp. 1–63" source_date: 1993 source_tier: 1
Quantitative claim, Tier 1 required and met. This confirms claim-patriot-army-effectiveness-claim-collapsed-from-96-to-52-percent's 52%/40% figures, and improves on that note's own source_quote: the existing note joins a phrase from the abstract ("...current claim that Patriot destroyed 52 percent of the Scud warheads that it engaged") to the 40%-highest-confidence clause with an ellipsis, implying they are one continuous sentence. In the primary text they are two separate sentences in two separate locations (abstract vs. introduction, p. 2) that both state the 52% figure independently; the introduction's version above is the single unbroken sentence that actually carries both the 52% and 40% figures together. Recommend the existing note's source_quote be corrected to this single-sentence form at next promotion/audit pass.
Claim: Lewis and Postol (1993) state that the Army's wartime claim was that Patriot successfully intercepted 45 of 47 Scuds, or 96 percent — the baseline the 52% figure revises downward from
The introduction states: "During the war, and in the weeks that followed, the US Army maintained that Patriot had successfully intercepted almost all of the Scuds (45 out of 47, or 96 percent). Since then, the Army has several times lowered its claims of Patriot effectiveness, most recently to just below 60 percent overall — over 40 percent in Israel and over 70 percent in Saudi Arabia."
"the US Army maintained\nthat Patriot had successfully intercepted almost all of the Scuds (45 out of47,\nor 96 percent)." — exact string (line breaks and missing "of47" space preserved) confirmed
grounded: trueviaquote_check.
source_quote: "the US Army maintained that Patriot had successfully intercepted almost all of the Scuds (45 out of 47, or 96 percent)." source_url: "https://scienceandglobalsecurity.org/archive/sgs04lewis.pdf" source_sha: "51b2ffeee1f0c11ce5997a24eb9aefccb3957da265efac4b1a5413678bba4a48" source_title: "Video Evidence on the Effectiveness of Patriot during the 1991 Gulf War" source_author: "George N. Lewis and Theodore A. Postol" source_venue: "Science & Global Security, 1993, Volume 4, pp. 1–63" source_date: 1993 source_tier: 1
Quantitative claim, Tier 1 required and met. This is the wartime baseline already embedded in claim-patriot-army-effectiveness-claim-collapsed-from-96-to-52-percent's title ("fell from a wartime 96% to a later 52%") but not previously carried as its own quoted, independently-verified figure; confirming it closes the last unverified number in that note's headline claim. Note also — not previously in the vault — that the paper separately reports a different Army metric, "just below 60 percent overall" successful-intercept rate (broken out by country), which is not the same figure as the 52%-warhead-destruction claim; see Further leads.
Answer to the central question
Yes. Both quoted claims in the topic question are verbatim in Lewis & Postol (1993), confirmed by direct primary-source read and quote_check against the exact extracted text (Tier 1, tls: verified). The 52%/40% figure in the existing vault note is accurate in substance but is constructed by joining two separate sentences from two different parts of the paper; a cleaner single-sentence source of the same two figures exists in the introduction and is recorded above. question-verify-lewis-postol-1993-patriot-video-evidence-quotes can be marked answered/verified-verbatim at promotion.
Further leads
- The Army's "just below 60 percent overall" successful-intercept claim (over 40% in Israel, over 70% in Saudi Arabia) is a distinct metric from the 52%-warhead-destruction figure — intercept success vs. warhead destruction are not the same claim — and is not yet captured anywhere in the vault; a future note could separate these two Army metrics explicitly.
- Footnotes 2 and 3 (attached to "warhead kills" and "'highest confidence' category") were not traced in this session — they likely cite the specific Army/DoD briefing or classified data the 52%/40% figures come from and could sharpen the claim-army-tsm-spreadsheet-and-brl-survey-flaws-undermined-patriot-kill-claims thread.
- Box A (p. 11) is described in-text as directly refuting the argument that video "detonation location uncertainty" prevents determining hits/misses from press footage — a candidate for a distinct technical-mechanism claim about the paper's own methodology, not pulled into this capture.
- The paper states its video sample covers "33 intercept attempts contained in 18 engagements," about 38 percent of all intercept attempts during the war — a sampling-coverage figure relevant to how representative the video-based finding is, not captured here.
Safety flags
None. The primary source is a scanned 1993 peer-reviewed physics-and-security journal article — third-person academic prose throughout, no addressed-to-AI language, override language, claimed authority, tier self-assignment, file-system instructions, credential requests, or urgency framing. extract_pdf reported tls: "verified", so no weak-transport elevation applies.
Entity candidates
- US Army (Patriot program office / DoD briefers) — institution — the older, institutional claimant whose 96%→52%→40%-high-confidence figures this entire paper measures itself against; the paper's whole rhetorical structure is "here is what they claimed, here is what the public record shows instead." This is the foundational-figure-on-the-other-side of the ancestry this capture traces, and per the standing instruction should be flagged before the analysts who contest it.
- George N. Lewis — person — co-author; already has entity-george-n-lewis in the vault.
- Theodore A. Postol — person — co-author; MIT Program in Science, Technology, and Society; not yet confirmed present as its own entity page in this session — check before creating.
- Raytheon Company — organization — supplied a collection of news-broadcast videos to Congress that the authors analyzed as part of their video corpus; manufacturer of the Patriot system itself, an interested party providing evidence used against the Army's own claims.