Does a Tier 1–2 primary (RAND personnel/project records, Isaacs's own writing, or Nash's or Blackwell's own testimony) confirm Rufus Isaacs, John Nash, and David Blackwell were direct RAND colleagues, not just co-listed alumni of different eras?
The 2026-08-22 capture 10-inbox/raw/2026-08-22-hop-isaacs-rand-precursor-bridge.md and the claim-note it produced, claim-isaacs-rand-colleagues-nash-blackwell-confirmed-overlap, rest the Isaacs–Nash–Blackwell RAND-colleague bridge on three tertiary/secondary sources: Wikipedia's colleague roster for Isaacs (Tier 4), INFORMS's biographical profile of Isaacs (Tier 3), and MacTutor's biography of Blackwell (Tier 3). Blackwell's RAND tenure ("three summers between 1948 and 1950") is independently dated and does overlap Isaacs's stated 1948–1955 window, which raises plausibility — but per 00-meta/specs/sources.md's historical-claim escalation rule, a load-bearing bridge claim like this one needs a Tier 1–2 source, and none has been read.
What would answer this: a RAND institutional history (e.g. the Jardini dissertation on RAND's early mathematics/economics divisions, or RAND's own published project rosters from 1948–1955); Isaacs's own writing, memoir, or interview naming Nash or Blackwell directly; Nash's or Blackwell's own writing or a documented interview in which either names Isaacs or describes working alongside him; or a scholarly history of RAND's game-theory group that lists a shared project, division, or seminar rather than just an institutional affiliation.
Candidate next moves: check whether Nash's own 1952 RAND report (D-1164, already read directly for claim-nash-1952-rand-report-confirms-hex-proof-non-constructive) names any colleagues in its acknowledgments; check Blackwell's MacTutor biography's own citations for a fuller RAND-years account; search for a published RAND organizational history covering the mathematics division circa 1948–1955.
2026-08-23 progress: Both candidate primaries above were located and read in full. Nash's D-1164 names only J. D. Williams (RAND math-division head, who requested the report) and Claude Shannon (a conversation credited as inspiration) as contemporary contacts — not Isaacs or Blackwell (claim-nash-1952-rand-report-names-williams-shannon-not-isaacs-or-blackwell). Blackwell's own 2002–2003 Bancroft Library oral history, read start to finish across two interview sessions devoted to his RAND years, names a dozen RAND-era colleagues but never Isaacs or Nash (claim-blackwell-1948-1950-rand-oral-history-omits-isaacs-and-nash). The Wikipedia colleague-roster sentence was also traced to its source: it carries no inline citation, and neither of the Wikipedia article's two listed Biography references (Yu 1979, Breitner 2005) contains it on inspection (claim-wikipedia-isaacs-colleague-roster-uncited-untraceable-to-its-references). Two negative primaries narrow the claim but do not close the question — an interview and a report can each omit a real acquaintance nobody asked about. Remaining unread leads, all attempted and blocked this session: Isaacs's own first RAND paper (Games of Pursuit, P-257, 1951 — rand.org 403s every tooling route, no Wayback Machine capture located); DeGroot's 1986 published interview of Blackwell in Statistical Science (Project Euclid returned an Incapsula bot-detection block); Isaacs's own 1969 JOTA retrospective essay "Differential Games: Their Scope, Nature, and Future" (Springer paywalled); MacTutor's own Isaacs biography (surfaced in search, not yet read). Status remains open.
2026-09-17 progress: P-257 was found and read in full this session — the earlier "no Wayback Machine capture located" note above was a wrong-guess artifact, not a genuine block: the PDF sits at .../pubs/papers/2006/P257.pdf, one folder-number off from the 2008 path guessed on 2026-08-22. The paper names no RAND colleague at all, Isaacs's own longtime associates included, but it is a solo technical note with no acknowledgments section of any kind, so its silence carries less weight than D-1164's or the oral history's targeted omissions (claim-isaacs-1951-rand-paper-names-no-colleagues). The other two remaining leads were re-attempted and remain genuinely blocked, not merely unattempted: the 1969 JOTA retrospective's abstract is free but its body sits behind a Springer institutional login, and a parallel Internet Archive microfilm holding of the same issue has no readable file; DeGroot's 1986 interview is still behind the identical Incapsula bot-detection wall found in 2026-08-23, confirmed again on two independent fetch routes this session. Three candidate primaries are now exhausted without a positive result; the two still-unread ones (JOTA retrospective, DeGroot interview) are the only plausible remaining paths to a confirmation. Status remains open.