V2 TO V3 — EXACT PROSE COMPARISON

v2 4750 words; v3 5199 words; net +449.
Section-wise alignment: 4390 retained (92.42% of v2), 360 removed, 809 added.
Includes headings, revision notes, tables and references; excludes equation tokens and TeX commands. Gross counts depend on the stated alignment.

SECTION | V2 | V3 | NET | REMOVED | ADDED
Title, abstract, and revision note | 492 | 596 | 104 | 17 | 121
1. Introduction | 457 | 463 | 6 | 12 | 18
2. Related Work | 276 | 276 | 0 | 0 | 0
3. Model | 1350 | 1403 | 53 | 100 | 153
4. Results | 958 | 1228 | 270 | 209 | 479
5. Conclusion | 547 | 549 | 2 | 20 | 22
6. Notes and References | 670 | 684 | 14 | 2 | 16

Original February preprint: 3,055 prose words. May v2: 4,750. Rejected broad November draft: 4,948. Final minimal v3: 5199.
The rejected draft is not the release. Its ordered-core and new adjustment-model architecture is absent from final v3.
This is a substantive formal repair, not a new economic framing. NR adds an explicit payoff restriction. The new proof keeps BPW and uses strong Nash as a sufficient route.
Outside the mathematical sections, changes align theorem summaries, add the v3 date/note, and adjust the final methodological note. No unsolicited price-appreciation disclaimer was added.
The full change ledger and primary-source reasoning are in paper1-change-review.txt and paper1-evidence.json.