The Maker-Breaker directed triangle game
Pith reviewed 2026-05-18 07:46 UTC · model grok-4.3
The pith
In the Maker-Breaker directed triangle game on tournaments, Breaker wins on parity boards with fewer than 7 vertices but Maker wins for larger sizes, with a bias threshold of order sqrt(n).
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
For the parity tournament on n vertices, Breaker has a winning strategy for 3 ≤ n < 7 while Maker wins for n ≥ 7; the (1:b) bias threshold satisfies sqrt((1/12+o(1))n) ≤ b*(n) ≤ sqrt((8/3+o(1))n); Maker wins the game on T(n,p) with probability tending to 1 for fixed p in (0,1); and the flip-bias threshold kappa*(n) is of order n^2 with explicit bounds like n(n-11)/12 ≤ kappa*(n) ≤ (n^2-1)/8 for odd n ≥ 11.
What carries the argument
The parity tournament on which the directed triangle hypergraph is defined, serving as a concrete board to determine the exact threshold for the game.
If this is right
- The bias threshold has the same sqrt(n) order as the classical undirected Maker-Breaker triangle game.
- Maker can force a directed triangle in random tournaments regardless of the fixed direction probability p in (0,1).
- The number of edge flips Breaker needs to win is on the order of n squared for the parity tournament.
Where Pith is reading between the lines
- Direction constraints in the triangle game do not alter the asymptotic bias threshold compared to the undirected case.
- Similar sharp thresholds might exist for other specially constructed tournaments beyond the parity one.
- The flip-bias model could be applied to study robustness of Maker-Breaker games under small perturbations to the board.
Load-bearing premise
The game is played on a tournament, meaning exactly one directed edge between every pair of vertices, and the winning sets are the directed 3-cycles in that tournament.
What would settle it
An explicit winning strategy for Maker in the unbiased game on the parity tournament with 6 vertices, or for Breaker on the parity tournament with 8 vertices.
read the original abstract
In this work, we investigate Maker-Breaker directed triangle games, a directionally constrained variant of the classical Maker-Breaker triangle game. Our board of interest is a tournament, and the winning sets are all $3$-cycles present in the tournament. We begin by studying the Maker-Breaker directed triangle game played on a specially defined tournament called the parity tournament, and we identify the board size threshold to be $n=7$, which is to say that for a parity tournament on $n$ vertices, Breaker has a winning strategy for $3\le n<7$, while Maker can ensure a win for herself for $n\ge7$. For the $(1:b)$ biased version of this game, we prove that the bias threshold $b^*(n)$ satisfies $\sqrt{\left(1/12+o(1)\right)n}\le b^*(n)\le\sqrt{\left(8/3+o(1)\right)n}$, which matches the order of magnitude, namely $\sqrt n$, of the bias threshold for the undirected counterpart of this game. Next, we consider the game on random tournaments $T(n,p)$, wherein the edge between $i$ and $j$, for each $i<j$, is directed from $i$ to $j$ with probability $p$, independent of all else. We prove that Maker wins this game with probability tending to $1$ as $n\to\infty$ for any fixed $p\in(0,1)$. Extending the notion of bias from undirected games to our directed framework, we introduce the flip-biased Maker-Breaker directed triangle game on the parity tournament with flip budget $\kappa(n)$, where Breaker may strategically flip the directions of at most $\kappa(n)$ edges before the game begins. We show that the flip-bias threshold $\kappa^*(n)$ for this game is of order $n^2$. More precisely, for odd $n\ge11$ we show $n(n-11)/12\le\kappa^*(n)\le (n^2-1)/8$, and for even $n\ge14$ we show $n(n-14)/12+1\le\kappa^*(n)\le n^2/8+n/4-1$.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper investigates Maker-Breaker directed triangle games on tournament boards, with winning sets being the directed 3-cycles. It establishes an exact threshold of n=7 for the parity tournament (Breaker wins for 3 ≤ n < 7, Maker wins for n ≥ 7). It proves that the (1:b) bias threshold satisfies √((1/12+o(1))n) ≤ b*(n) ≤ √((8/3+o(1))n). It shows Maker wins with probability tending to 1 on random tournaments T(n,p) for fixed p ∈ (0,1). It introduces the flip-biased variant on the parity tournament and proves that the flip-bias threshold κ*(n) is of order n², with explicit bounds n(n-11)/12 ≤ κ*(n) ≤ (n²-1)/8 for odd n ≥ 11 and n(n-14)/12 + 1 ≤ κ*(n) ≤ n²/8 + n/4 - 1 for even n ≥ 14.
Significance. If the results hold, the work meaningfully extends Maker-Breaker theory to directed tournaments. The exact n=7 threshold for the parity construction, the matching √n order for the bias threshold with the undirected case, the high-probability win on T(n,p), and the concrete linear factors in the flip-bias bounds (e.g., n-11 and n-14 arising from triangle counting) are all strengths. The paper supplies explicit strategies, pairing arguments, and probabilistic calculations that support the central claims without reducing to fitted parameters.
minor comments (4)
- [§2] §2 (parity tournament): the explicit pairing strategy for Breaker when 3 ≤ n < 7 could include a small diagram or table listing the pairs for n=5 and n=6 to improve readability.
- [§3] §3 (bias threshold): the o(1) terms in the lower and upper bounds on b*(n) are stated but the error terms from the degree counting could be bounded more explicitly in the proof of the upper bound.
- [§4] §4 (random tournaments): the application of the random-hypergraph criterion is standard, but the precise probability threshold used for the appearance of directed triangles should be cross-referenced to the relevant lemma.
- [§5] Notation: the flip budget is denoted κ(n) while the threshold is κ*(n); ensure consistent use of the asterisk throughout the flip-bias section.
Simulated Author's Rebuttal
We thank the referee for the positive and accurate summary of our results on the Maker-Breaker directed triangle game, including the exact threshold for the parity tournament, the bias threshold bounds, the result on random tournaments, and the flip-bias analysis. We appreciate the recommendation for minor revision.
Circularity Check
No significant circularity identified
full rationale
The paper derives its threshold results (exact n=7 switch for the parity tournament, sqrt(n)-order bias bounds, whp Maker win on T(n,p), and n^2-order flip-bias bounds) via explicit Breaker/Maker strategies, direct counting of directed triangles per vertex/edge in the parity construction, pairing/degree arguments, and standard random-hypergraph criteria. None of these steps reduce a claimed prediction or first-principles result to a fitted input or self-citation by the paper's own equations; the concrete factors (n-11, n-14) and o(1) terms arise from independent enumeration inside the parity tournament definition. The derivation chain is therefore self-contained against the stated board, winning sets, and game rules.
Axiom & Free-Parameter Ledger
axioms (2)
- standard math A tournament is a directed complete graph with exactly one directed edge between every pair of vertices.
- domain assumption The winning sets are all directed 3-cycles present in the tournament.
Lean theorems connected to this paper
-
IndisputableMonolith/Foundation/RealityFromDistinction.leanreality_from_one_distinction unclear?
unclearRelation between the paper passage and the cited Recognition theorem.
Theorem 2.1 … board size threshold … n=7 … parity tournament … cycle hopping strategy
-
IndisputableMonolith/Cost/FunctionalEquation.leanwashburn_uniqueness_aczel unclear?
unclearRelation between the paper passage and the cited Recognition theorem.
bias threshold … √((1/12+o(1))n) … √((8/3+o(1))n)
What do these tags mean?
- matches
- The paper's claim is directly supported by a theorem in the formal canon.
- supports
- The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
- extends
- The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
- uses
- The paper appears to rely on the theorem as machinery.
- contradicts
- The paper's claim conflicts with a theorem or certificate in the canon.
- unclear
- Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.