Verify Nash's Hex strategy-stealing proof (non-constructive; PSPACE-hard to construct) against a math-history primary or scholarly secondary

claim-nash-hex-first-player-win-proof-is-non-constructive rests on Tier-4 Wikipedia ("Strategy-stealing argument") for its load-bearing point: that Nash's classic proof of first-player win in Hex is non-constructive, and that computing an explicit winning strategy was later shown PSPACE-hard (attributed to Even and Tarjan, 1976). Under the sources.md escalation rule, a claim whose entire interest rests on one non-obvious fact should not sit on a tertiary source alone.

What to read / pull:

• A published account of Nash's own unpublished Hex work — histories of game theory sometimes reproduce the RAND-era note or oral-history testimony (e.g. Sylvia Nasar's A Beautiful Mind, or game-theory histories covering Nash and Hex/"Nash" the game).
• Even, S. and Tarjan, R. E., "A Combinatorial Problem Which Is Complete in Polynomial Space" (1976) — the primary for the PSPACE-hardness claim, read directly rather than via Wikipedia's summary.
• Martin Gardner's "Hex, the Game with the Beautiful Idea" or a comparable math-popularization piece known for accurately relaying Nash's proof.

What it gates: moving the note off seedling / clearing its [unverified-source] flag. Also underpins the three-way Hex thread linking claim-shannon-moore-1950-analog-hex-machine-move-as-saddle-point and claim-gale-1979-hex-draw-impossibility-equivalent-to-brouwer-fixed-point.