/- Copyright 2026 The Formal Conjectures Authors. Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at https://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. -/ module public import FormalConjecturesUtil

Erdős Problem 1150

References:

    erdosproblems.com/1150

    [OAI26a] OpenAI, Asymptotically minimal maxima of real Littlewood polynomials. OpenAI Math Release preprint (2026). https://github.com/openai/math/blob/adc7f1241b42e322a6451854ab7e4b4c146bf78a/preprints/Asymptotically-minimal-maxima-of-real-Littlewood-polynomials-September-23-2026/paper.pdf

@[expose] public sectionopen scoped Polynomialnamespace Erdos1150

Is there some constant $c > 0$ such that, for all large enough $n$ and all polynomials $P$ of degree $n$ with coefficients in ${-1, 1}$, $$\max_{|z|=1} |P(z)| > (1 + c) \sqrt{n}?$$

No. This was disproved by an internal OpenAI model in September 2026 [OAI26a]; see the linked Lean proof.

@[category research solved, question, AMS 12 30, formal_proof using lean4 at "https://github.com/openai/math/blob/adc7f1241b42e322a6451854ab7e4b4c146bf78a/lean/OAI/Analysis/Littlewood/Main.lean#L187"] theorem erdos_1150 : ¬ ∃ c > 0, ∀ᶠ n in Filter.atTop, ∀ P : ℂ[X], (∀ i ≤ P.natDegree, P.coeff i = - 1 ∨ P.coeff i = 1) → P.natDegree = n → ⨆ z : Metric.sphere (0 : ℂ) 1, ‖P.eval (z : ℂ)‖ > (1 + c) * Real.sqrt n := ⊢ ¬∃ c > 0, ∀ᶠ (n : ℕ) in Filter.atTop, ∀ (P : ℂ[X]), (∀ i ≤ P.natDegree, P.coeff i = -1 ∨ P.coeff i = 1) → P.natDegree = n → ⨆ z, ‖Polynomial.eval (↑z) P‖ > (1 + c) * √↑n All goals completed! 🐙

The trivial lower bound from Parseval's identity: for any polynomial $P$ of degree $n$ with coefficients in ${-1, 1}$, we have $\max_{|z|=1} |P(z)| \geq \sqrt{n+1}$.

This follows from Parseval's identity: $$\frac{1}{2\pi} \int_0^{2\pi} |P(e^{i\theta})|^2 d\theta = \sum_{k=0}^{n} |a_k|^2 = n+1$$ since each $|a_k|^2 = 1$. The circle average is bounded by the pointwise supremum squared, so $\max_{|z|=1} |P(z)|^2 \ge n+1$, whence $\max_{|z|=1} |P(z)| \ge \sqrt{n+1}$.

P:ℂ[X]n:ℕhcoeff:∀ i ≤ P.natDegree, P.coeff i = -1 ∨ P.coeff i = 1hdeg:P.natDegree = nN:ℕ := n + 1hN:N = n + 1hcoeff_norm:∀ i ∈ Finset.range N, ‖P.coeff i‖ = 1hsupport:P.support = Finset.range NhParseval:Real.circleAverage (fun z ↦ ‖Polynomial.eval z P‖ ^ 2) 0 1 = ↑Nhdeg':P.natDegree < Nhtri:∀ z ∈ Metric.sphere 0 1, ‖Polynomial.eval z P‖ ≤ ↑Nhbdd:BddAbove (Set.range fun z ↦ ‖Polynomial.eval (↑z) P‖)S:ℝ := ⨆ z, ‖Polynomial.eval (↑z) P‖hSdef:S = ⨆ z, ‖Polynomial.eval (↑z) P‖hS_nonneg:0 ≤ Shbound_sq:∀ z ∈ Metric.sphere 0 1, ‖Polynomial.eval z P‖ ^ 2 ≤ S ^ 2hCI:CircleIntegrable (fun z ↦ ‖Polynomial.eval z P‖ ^ 2) 0 1h_avg_le:↑N ≤ S ^ 2h_sqrt:√↑N ≤ S⊢ S ≥ √(↑n + 1) calc Real.sqrt ((n : ℝ) + 1) = Real.sqrt ((N : ℝ)) := P:ℂ[X]n:ℕhcoeff:∀ i ≤ P.natDegree, P.coeff i = -1 ∨ P.coeff i = 1hdeg:P.natDegree = nN:ℕ := n + 1hN:N = n + 1hcoeff_norm:∀ i ∈ Finset.range N, ‖P.coeff i‖ = 1hsupport:P.support = Finset.range NhParseval:Real.circleAverage (fun z ↦ ‖Polynomial.eval z P‖ ^ 2) 0 1 = ↑Nhdeg':P.natDegree < Nhtri:∀ z ∈ Metric.sphere 0 1, ‖Polynomial.eval z P‖ ≤ ↑Nhbdd:BddAbove (Set.range fun z ↦ ‖Polynomial.eval (↑z) P‖)S:ℝ := ⨆ z, ‖Polynomial.eval (↑z) P‖hSdef:S = ⨆ z, ‖Polynomial.eval (↑z) P‖hS_nonneg:0 ≤ Shbound_sq:∀ z ∈ Metric.sphere 0 1, ‖Polynomial.eval z P‖ ^ 2 ≤ S ^ 2hCI:CircleIntegrable (fun z ↦ ‖Polynomial.eval z P‖ ^ 2) 0 1h_avg_le:↑N ≤ S ^ 2h_sqrt:√↑N ≤ S⊢ √(↑n + 1) = √↑N P:ℂ[X]n:ℕhcoeff:∀ i ≤ P.natDegree, P.coeff i = -1 ∨ P.coeff i = 1hdeg:P.natDegree = nN:ℕ := n + 1hN:N = n + 1hcoeff_norm:∀ i ∈ Finset.range N, ‖P.coeff i‖ = 1hsupport:P.support = Finset.range NhParseval:Real.circleAverage (fun z ↦ ‖Polynomial.eval z P‖ ^ 2) 0 1 = ↑Nhdeg':P.natDegree < Nhtri:∀ z ∈ Metric.sphere 0 1, ‖Polynomial.eval z P‖ ≤ ↑Nhbdd:BddAbove (Set.range fun z ↦ ‖Polynomial.eval (↑z) P‖)S:ℝ := ⨆ z, ‖Polynomial.eval (↑z) P‖hSdef:S = ⨆ z, ‖Polynomial.eval (↑z) P‖hS_nonneg:0 ≤ Shbound_sq:∀ z ∈ Metric.sphere 0 1, ‖Polynomial.eval z P‖ ^ 2 ≤ S ^ 2hCI:CircleIntegrable (fun z ↦ ‖Polynomial.eval z P‖ ^ 2) 0 1h_avg_le:↑N ≤ S ^ 2h_sqrt:√↑N ≤ S⊢ √(↑n + 1) = √(↑n + 1); All goals completed! 🐙 _ ≤ S := h_sqrtend Erdos1150