Smale's mean value constant: is it exactly 4?
For a complex polynomial of degree with and , let , and let be the supremum of over all such of degree . Smale (1981) proved for every and asked whether can be replaced by ; that was disproved in 2026. What is the exact value of Smale's mean value constant ? Jatar and Ng (2026) list whether as open.
- Result
- Proved(see note)
- Status
- Candidate (review pending)
- AI contribution
- AI co-developed
- Method
- Construction
- Field
- Complex analysis
- Posed by
- Stephen Smale; the exact value restated as open by Aneesh Jatar and Tuen-Wai Ng (2026)
- Year posed
- 1981
- Years open
- 45y
- Solved
- 2026-11-06
- Model
- Claude; ChatGPT (versions not stated)
- Vendor
- Anthropic; OpenAI
- Collaborators
- Dave Goel
- Verification
- Lean-checked, statement unaudited
- Publication
- Preprint
- Significance
- 45 / 100
- Disclosed cost
- —
- Wikipedia
- No dedicated article
What was actually shown
Theorem 1.1: . Each is known to be below , so the supremum is not attained, and the proof gives no bound on the degree needed to exceed a given . The proof combines the Jatar-Ng reduction to admissible univalent-Blaschke pairs with an explicit construction. Lean: Smale.P030.Rows.thm_1_1 (and kstar_eq_four, tendsto_Kd_four) at tag v2.50. This answers the exact-value question; the conjecture was disproved earlier (see the related entry).
What the AI did
From the paper's AI disclosure and Hexagon provenance statement: the work was done overwhelmingly by AI models, steered by the author in a continuous critical dialogue, alternating between Anthropic's Claude and OpenAI's ChatGPT. The author takes full responsibility. Hexagon labels it a mix of human-written and AI-generated text.
Verification
No independent mathematician has checked this yet. Lean-checked on this site's ladder, not Lean-verified. Checked here on 15 November 2026 from tag v2.50 of github.com/david44hello/paper_smale-mean-value-constant-equals-4 (Lean v4.35.0, mathlib v4.35.0 pinned): 23 Lean files, about 7,000 lines; a grep for sorry, axiom, admit, native_decide, unsafe, implemented_by, extern, opaque, macro and elab finds nothing. The headline definitions were read line by line: Kd is the supremum in EReal, over polynomials of degree d with p(0)=0 and p'(0)=1, of the least |p(b)|/|b| over critical points b; Kstar is the supremum over d at least 2; thm_1_1 restates both in the paper's p'(0) nonzero normalization and proves the two agree. That matches the problem as Smale and Jatar-Ng pose it. Not rebuilt here and axioms not printed here; the README reports a Comparator run on 9 November with no public log or CI. Formal Conjectures states only the K = 1 form, so nothing independent anchors the K* = 4 statement; a Palomar listing (PALOMAR-2026-11-12-000015) is cited by Hexagon. The paper's argument was read in outline only.
Sources
- PaperExpanded version, Hexagon 2611.00283
- Lean proofLean 4 proof, tag v2.50 (headline thm_1_1 in Smale/P030/Rows/Thm11.lean)
- Lean statementHeadline definitions Kd and Kstar (Smale/Defs/Basic.lean)
- Problem recordJatar and Ng, Smale's mean value constant and admissible univalent Blaschke pairs (Section 9 lists K* = 4 as open)
- OtherHexagonRelated entry: Smale's mean value conjecture (K = 1), disproved
Submitted by MellowGander562 on