\documentclass[11pt]{amsart} \usepackage[margin=1in]{geometry} \usepackage{amsmath,amssymb,amsthm} \usepackage[hidelinks]{hyperref} \usepackage{booktabs} \newtheorem{theorem}{Theorem} \newtheorem{corollary}[theorem]{Corollary} \theoremstyle{remark} \newtheorem{remark}[theorem]{Remark} \title[A counterexample of degree 29\,440 to Smale's mean value problem]{An explicit counterexample of degree 29\,440\\ to Smale's mean value problem with constant one} \author{David Niedbala Giraudin} \address{Meaux, France} \email{mrjack77100@gmail.com} \thanks{ORCID 0009-0009-1526-1178.} \date{November 7, 2026} \subjclass[2020]{30C15, 30C10, 65G20} \keywords{Smale's mean value problem, critical points of polynomials, interval arithmetic, Krawczyk test} \begin{document} \sloppy \begin{abstract} For a complex polynomial $p$ with $p(0)=0$ and $p'(0)=1$, let $S(p,0)=\min_{p'(c)=0}|p(c)/c|$. Adamczewski and Colbrook recently gave an explicit polynomial of degree $29\,984$ with $S(p,0)>1$, certified by interval arithmetic, built on the counterexample found by GPT-6 Astra. Within their family of polynomials, with re-chosen coefficients, we give an explicit $p\in\mathbb{Q}(i)[z]$ of degree $29\,440$ with \[ 1.00000093069254310 < S(p,0) < 1.00000093069254311 . \] Jatar and Ng have shown that Smale's constant $\sup_p S(p,0)$ exceeds $1.1018$, through polynomials of unspecified degree; the point here is an explicit polynomial with a certified bound. The bound is certified twice: by a ball-arithmetic Krawczyk certificate covering all $29\,439$ critical points, and, independently, by the Adamczewski--Colbrook verifier, which confirms $29\,418$ of them and is inconclusive (insufficient working precision) on the other $21$. The exact coefficients and both verification programs are provided as ancillary files. \end{abstract} \maketitle \section{Introduction} Let $p$ be a complex polynomial of degree $n\ge 2$ and $a\in\mathbb{C}$ with $p'(a)\neq 0$. Smale \cite{Smale} proved that \[ S(p,a)=\min_{p'(c)=0}\left|\frac{p(c)-p(a)}{(c-a)\,p'(a)}\right|\le 4 \] and asked whether the constant $4$ can be replaced by $1$, or even by $(n-1)/n$. The constant-one version was disproved by GPT-6 Astra, with a formal Lean~4 proof \cite{Ad}. Adamczewski and Colbrook \cite{AC} explained the underlying mechanism and gave an explicit polynomial $p\in\mathbb{Q}(i)[z]$ of degree $29\,984$ with $S(p,0)>1.00000049595801$, certified by interval arithmetic \cite{ACcode}. Independently, Jatar and Ng \cite{JN} reduced the determination of Smale's mean value constant $\mathcal M=\sup S(p,0)$, the supremum over normalized polynomials, to an extremal problem for admissible pairs of a univalent function and a finite Blaschke product, and showed that $\mathcal M>1.1018$; their polynomials come from an approximation argument, and no explicit degree is given. Thus the value $1.00000093\ldots$ below is not a bound on $\mathcal M$; what is new here is only the degree of an explicit, certified example. This note records a smaller explicit counterexample. It uses \emph{the same construction} as \cite[Appendix~A]{AC}: the same lemniscate coordinate, model anchor, Blaschke parameter, correction centres and correction structure. Only the truncation level $N$, the size of the high-degree tail and the coefficient tables are changed. The reduction in degree is modest (about $1.8\%$); no new mechanism is claimed. \section{The polynomial} We keep the notation of \cite[Appendix~A]{AC}: $m=16$, $A=1061603/250000$, $S=A+1/A$, $q(z)=(z^{16}+iA)/S$, the model anchor $a$, the Blaschke parameter $r$ and the truncated model $h_L$ of \cite[(A.5)]{AC}. We take \[ N=1840,\qquad L=N-2=1838,\qquad \text{tail blocks } t=48, \] and set \begin{align*} P(z)&=(z-a)\frac{h_L(z)}{h_L(a)}+\Delta_0(z)+X(z)+H(z),\\ \Delta_0&=(z-a)^2\Bigl(K R_0+\sum_{j=0}^{3}\phi_j(z-\mu)^j\Bigr),\qquad K=(z-\xi_1)^2(z-\xi_2)^2,\quad \mu=\tfrac{\xi_1+\xi_2}{2},\\ X&=(z-a)^2\sum_{\ell=0}^{127}\sum_{k=0}^{15}x_{\ell k}\Bigl(\frac{5z}{6}\Bigr)^{k}q(z)^{\ell},\\ H&=\omega\Bigl(\gamma\, q(z)^{N}+q(z)^{N-t}\sum_{\ell=0}^{t-1}\sum_{k=0}^{15}e_{\ell k}\Bigl(\frac{5z}{6}\Bigr)^{k}q(z)^{\ell}\Bigr), \end{align*} with $R_0$ the correction polynomial of \cite{AC} (its $1024$ coefficients re-chosen), rational $\omega$ with $|\omega|=1$ exactly, rational $\gamma>0$, and a rational base point $\hat a$. The normalized polynomial is \[ p(x)=\frac{P(\hat a+x)-P(\hat a)}{P'(\hat a)} . \] All data are exact elements of $\mathbb{Q}$ or $\mathbb{Q}(i)$ and are listed in the ancillary file \texttt{polynomial\_N1840.json}, which follows the format of \cite{ACcode} (field \texttt{defining\_formula}). In particular $p\in\mathbb{Q}(i)[z]$. Only $\gamma q^N$ reaches degree $16N=29\,440$; all other terms have degree at most $29\,439$, and the leading coefficient $\omega\gamma/S^{N}$ is nonzero. Hence $\deg p=29\,440$, and $p(0)=0$, $p'(0)=1$ by construction. \begin{theorem}\label{thm:main} The polynomial $p\in\mathbb{Q}(i)[z]$ above has degree $29\,440$, its $29\,439$ critical points are simple, and \[ 1.00000093069254310 < S(p,0) < 1.00000093069254311 . \] In particular $p$ is a counterexample to Smale's mean value conjecture with constant one. \end{theorem} \begin{remark} Adamczewski and Colbrook note \cite[\S1.1]{AC} that perturbations $\varepsilon z^n$ of an explicit counterexample, together with Rouch\'e's theorem, give counterexamples in every larger degree. Applied to $p$, the same argument would give counterexamples in every degree $n\ge 29\,440$; we do not repeat it here, and Theorem~\ref{thm:main} only concerns degree $29\,440$. \end{remark} \section{Verification} Since $S(p,0)$ is unchanged by the affine normalization, it suffices to work with $P$ and $\hat a$: $S(p,0)=\min_{P'(c)=0}|Q(c)|$ with $Q(c)=\dfrac{P(c)-P(\hat a)}{(c-\hat a)P'(\hat a)}$. \subsection*{Certificate 1 (ball arithmetic)} Arithmetic is done with complex balls (Arb, via \texttt{python-flint}) at $1280$ bits, starting from the exact rationals of the specification. For each of $29\,439$ approximate critical points: \begin{enumerate} \item three Newton steps for $P'$ give a centre $z$; \item on the square box $D$ of half-side $\rho$ about $z$, an enclosure of $P''(D)$ is obtained from $P''(z)$ and a Cauchy bound $\sup|P'''|\le 6B/(R-\rho\sqrt2)^3$, where $B$ is a rigorous majorant of $|P|$ on the disc of radius $R$ about $z$ (term-by-term triangle inequality); \item the Krawczyk operator $K(D)=z-P'(z)/P''(z)+\bigl(1-P''(D)/P''(z)\bigr)(D-z)$ is shown to lie in the interior of $D$, which proves that $D$ contains exactly one zero of $P'$, and that it is simple; \item $|Q|$ is bounded below on all of $D$ by a second-order Taylor enclosure of $P$ with the same remainder bound, and a point counts as certified only if this lower bound exceeds $1$. \end{enumerate} All $29\,439$ points pass. The boxes have half-side at most $7.4\cdot10^{-40}$, the refined centres differ from the supplied ones by at most $7.7\cdot10^{-12}$, and any two supplied centres are at distance at least $6.05\cdot10^{-5}$; hence the boxes are pairwise disjoint. They thus contain $29\,439=\deg P'$ distinct zeros of $P'$, i.e.\ all critical points, each simple. The smallest certified value is \[ \min_{P'(c)=0}|Q(c)|\ \ge\ 1.0000009306925431017438\ldots, \] attained in a single box, which also gives the upper bound of Theorem~\ref{thm:main}. \subsection*{Certificate 2 (Adamczewski--Colbrook verifier)} The program \texttt{verify.cpp} of \cite{ACcode} (outward-rounded interval bounds with MPFR-directed endpoints, Rouch\'e test on critical disks) was recompiled with \texttt{N\_PARAM=1840}, \texttt{TAIL\_BLOCKS\_PARAM=48}, the expected parameters in \texttt{make\_parameters.py} being read from the specification; the specification passes all its exact-rational consistency checks. Run disk by disk on the $29\,439$ critical disks, it certifies $|Q|>1$ on $29\,418$ of them, with smallest lower bound $1.000000930692128\ldots$, on the same critical point as Certificate~1. On the remaining $21$ disks (indices, in the order of \texttt{critical\_disks\_N1840.csv}, listed in the ancillary summary) its fixed working precision does not suffice to conclude; these are covered by Certificate~1 only. \begin{table}[h] \centering \begin{tabular}{lcc} \toprule & \cite{AC} & this note\\ \midrule $N$ & $1874$ & $1840$\\ degree & $29\,984$ & $29\,440$\\ tail blocks & $8$ & $48$\\ certified lower bound for $S(p,0)$ & $1.00000049595801$ & $1.00000093069254310$\\ \bottomrule \end{tabular} \end{table} \section{Ancillary files} \begin{itemize} \item \texttt{polynomial\_N1840.json}: exact specification ($\mathbb{Q}$ and $\mathbb{Q}(i)$ data). \item \texttt{centres\_N1840.csv}: the $29\,439$ supplied critical-point centres, sorted by increasing $|Q|$. \item \texttt{verify\_ball\_N1840.py}: Certificate~1 for any range of critical points (Python with \texttt{python-flint} and \texttt{numpy}). \item \texttt{certificate\_N1840.log}: summary of both certificates. \item \texttt{critical\_disks\_N1840.csv}, \texttt{make\_parameters\_N1840.py}: input disks and the modified parameter script used to run the verifier of \cite{ACcode} (Apache License 2.0). \end{itemize} \section*{Status and acknowledgements} Theorem~\ref{thm:main} is a computer-assisted result: its proof is the pair of certificates above, which anyone can rerun from the ancillary files. The construction is entirely that of Adamczewski and Colbrook \cite{AC}, building on the counterexample found by GPT-6 Astra \cite{Ad}; only parameters were changed. The computations were carried out with the assistance of Claude (Anthropic). \begin{thebibliography}{9} \bibitem{Ad} T.~Adamczewski, \emph{Mean-value-problem: Lean 4 disproof of Smale's mean value conjecture ($K=1$)}, GitHub repository, 2026. \bibitem{AC} T.~Adamczewski and M.~J.~Colbrook, \emph{From boundary geometry to counterexamples in Smale's mean value problem}, arXiv:2610.36158, 2026. \bibitem{ACcode} M.~J.~Colbrook, \emph{Interval\_arithmetic\_counterexample\_Smale\_mean\_value}, GitHub repository, 2026. \bibitem{JN} A.~Jatar and T.~W.~Ng, \emph{Smale's mean value constant and admissible univalent Blaschke pairs}, arXiv:2610.21472, 2026. \bibitem{Smale} S.~Smale, The fundamental theorem of algebra and complexity theory, \emph{Bull. Amer. Math. Soc. (N.S.)} \textbf{4} (1981), 1--36. \end{thebibliography} \end{document}