Keyboard shortcuts

Press ← or → to navigate between chapters

Press ? to show this help

Press Esc to hide this help

Nonconvergence for all-odd subtraction games

Abstract

All-odd finite subtraction games have nonconvergent winning probabilities.

Definition 1.1 (Legal subtractions).

Formalization. D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.legalMoves (✓ std3).

Citation. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

Section 1, p. 2: “At a given position (number of chips in the pile) 0 ≤ m ≤ n, an element k′ ∈ {k ∈ S | k ≤ m} is randomly selected with equal probability, and the next player to move removes k′ chips from the pile.” The paper takes S to consist of positive integers. Here S is a finite set of natural numbers; the filter explicitly requires a positive subtraction and includes precisely those at most n.

Definition 1.2 (Probability of winning).

Formalization. D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.winProb (✓ std3).

Citation. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

Section 1, p. 2: “The first player who is confronted with a position m for which {k ∈ S | k ≤ m} = ∅, loses.” The function winProb S n is the probability that the next player wins. It is zero without a legal move and otherwise one minus the uniform average of the follower probabilities. The sum uses attach, whose elements carry their membership proofs; val is the underlying natural number. Subtraction of natural indices is truncated subtraction. The cardinality is cast to the reals, and the displayed quotient is real division. Recursion terminates because each legal subtraction is positive.

Definition 1.3 (The all-odd conjecture).

Formalization. D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.claim (✓ std3).

Citation. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

Section 3, p. 7: “However, we do not show that an all-odd subtraction set necesarily do not converge (although we conjeture it) nor we provide the exact limits of the even and odd subsequences for the (conjectured) non-convergent case.” The section assumes a finite nonempty subtraction set with gcd one. Odd natural numbers are positive, and nonconvergence means that no real limit exists.

Definition 1.4 (Alternating probability transform).

Formalization. D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.b (✓ std3).

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

The private definition b uses real powers and real arithmetic.

Definition 1.5 (Adjacent transformed probabilities).

Formalization. D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.d (✓ std3).

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

The private definition d adds consecutive values of b.

Lemma 1.6 (A newly legal subtraction).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.d_threshold (✓ std3). ∎

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

The private d_threshold identity includes the changing-denominator term b S n + 1. The cardinality is cast to the reals before adding one.

Lemma 1.7 (Before the first legal subtraction).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.d_initial (✓ std3). ∎

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

The private d_initial identity covers both possibilities for the next pile size.

Lemma 1.8 (Nonnegative adjacent sums).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.d_nonneg (✓ std3). ∎

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

The private d_nonneg theorem uses strong induction through changes of legal moves; it requires neither nonemptiness nor a gcd hypothesis.

Lemma 1.9 (The positive seed at the least subtraction).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.d_seed (✓ std3). ∎

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

The private d_seed theorem assumes membership and minimality of m. Natural subtraction in m - 1 is truncated.

Definition 1.10 (The finite adjacent-sum window).

Formalization. D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.invariant (✓ std3).

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

The private invariant definition sums over Finset.range s. The index N - s + r means truncated natural subtraction followed by addition.

Lemma 1.11 (Conservation after all moves are legal).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.invariant_step (✓ std3). ∎

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

The private invariant_step theorem applies when N bounds every element of S.

Lemma 1.12 (The conserved window at every later index).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.invariant_constant (✓ std3). ∎

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

The private invariant_constant theorem iterates conservation from any upper bound M.

Lemma 1.13 (The conserved window is at least two).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.invariant_positive (✓ std3). ∎

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

The private invariant_positive theorem evaluates the window at Finset.max’ S hne. Here hne is the proof of nonemptiness required by Finset.max’.

Theorem 1.14 (Nonconvergence without a gcd hypothesis).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.nonconvergence (✓ std3). ∎

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

The private nonconvergence theorem is the stronger statement consumed by result. Its hypotheses are exactly nonemptiness and oddness of every subtraction.

Theorem 1.15 (Every all-odd subtraction set fails to converge).

Proof. Machine-checked in Lean as D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.result (✓ std3). ∎

Resolves. Problems/capitelli-somma-2024-all-odd-nonconvergence (proved) by D5/S3/Combinatorics/Games/CapitelliSommaAllOddNonconvergence.result.

Source. Repository-derived.

Acknowledgement. Nicolás Capitelli, Francisco Somma (2024). On sequences arising from randomizing subtraction games. DOI: 10.48550/arXiv.2405.19593. URL: https://arxiv.org/abs/2405.19593v1.

Commentary.

Put b(n)=(-1)^n(1-2 winProb S n) and d(n)=b(n)+b(n+1). Oddness makes b satisfy a positive averaging recurrence. At a newly legal subtraction, the change in the averaging denominator contributes b(n)+1, which is nonnegative since b(n) lies between minus one and one. Strong induction gives d(n) nonnegative, with d(m-1)=2 at the first legal move m=min S. For N at least max S, the finite sum over s in S and r<s of d(N-s+r) is conserved by telescoping and is at least two. Convergence of winProb would force d, and hence this finite sum, to tend to zero. This contradiction proves the proposition. The argument also applies without the gcd hypothesis. It does not determine the even and odd subsequence limits or the multiplicities of the characteristic roots.

References