Bounded-size ω→-witness subtournament

Status partial medium confidence

Question 5.9 asks whether $\overrightarrow{\omega}(T)\geq k$ is always witnessed by a bounded-size subtournament (i.e., $f$ = identity). A 2026 paper by Crew, Fan, Koerts, Moore, and Spirkl (arXiv:2602.09863) proves a closely related weaker form (their Corollary 7): there exist two functions $f$ and $\ell$ such that $\overrightarrow{\omega}(T)\geq f(k)$ implies a subtournament of size $\leq\ell(k)$ with $\overrightarrow{\omega}\geq k$; this resolves the weaker Conjecture 5.8 of the source paper. Whether $f$ can be taken as the identity — the content of Question 5.9 — remains unclear from the available evidence.

Cited literature (1)

• partial Characterizing Large Clique Number in Tournaments (2026)
  Logan Crew, Xinyue Fan, Hidde Koerts, Benjamin Moore, Sophie Spirkl · arXiv preprint · arXiv:2602.09863

  Proves (Corollary 7) that there exist functions f and \ell such that every tournament with \overrightarrow{\omega}(T) \geq f(k) contains a subtournament of size \leq \ell(k) with \overrightarrow{\omega} \geq k, settling the weaker Conjecture 5.8 of the source paper; the paper self-reports answering a question of Aboulker--Aubian--Charbit--Lopes (2023), but Corollary 7 uses two functions rather than the single identity function required by Question 5.9.

Reviewer notes. arXiv:2602.09863 self-reports answering a question from Aboulker--Aubian--Charbit--Lopes (2023); their Corollary 7 establishes bounded-size witnesses for large (\geq f(k)) clique number, which is Conjecture 5.8, not the stronger identity-function form of Question 5.9. The source paper notes that Conjecture 5.10 (infinitely many k-\overrightarrow{\omega}-critical tournaments) would answer Question 5.9 negatively.

Auto-reviewed 2026-05-15 with claude-sonnet-4-6 (web search enabled).