Skip to content

Games equivalent as games and the fundamental lemma - #28

Merged
MarcIlunga merged 1 commit into
mainfrom
game-equivalence
Oct 13, 2026
Merged

MarcIlunga merged 1 commit into
mainfrom
game-equivalence

Conversation

@MarcIlunga

@MarcIlunga MarcIlunga commented Oct 12, 2026 •

Copy link
Copy Markdown
Collaborator

Proof-technique stack, layer 3 (games). Based on the H-coefficient PR, though independent of the two layers below it.

RandomSystem.GameEquivalent: two games equivalent as games, G ≡g H (CR18, Definition 4.16, p. 105). They give every transcript with every MBO unset the same probability, which is the cumulative condition CR18 states as equivalent to identical pre-winning behavior. Maurer 2002, Definition 5 (F^A ≡ G^C), is the same notion.

  • Won alike: GameEquivalent.winProbability_eq and Domain.Distinguisher.winProbability_eq_of_gameEquivalent (CR18, Lemma 4.15, p. 106; Maurer 2002, Lemma 7 and Lemma 6 (ii)).
  • The fundamental lemma: the advantage between the visible systems is at most the winning probability: GameEquivalent.abs_decisionProbability_sub_le and Domain.Distinguisher.advantage_le_winProbability_of_gameEquivalent (CR18, Lemma 4.16, p. 107; Maurer 2002, Lemma 7 and Theorem 1 (i)).
  • Common part: Distribution.abs_mass_sub_le_of_le: two laws of weight one above a common part ν differ on every event by at most 1 - ν.weight. The conditional-equivalence bound from Games with a monotone binary output: Games, DistinctionGames and the interface instance #4 now uses it instead of proving the step inline.

There is no ≡g notation: the systems level has no scoped notation of its own.

THEORY.md §1.5 gets a paragraph.

Checks: build with --wfail, axiom audit, linter set, anchors, math rendering, CR18 pages checked on the PDF.

The stack

Proof techniques for the HCTR2 port, one technique per PR, ordered by abstraction (merge bottom-up):

  1. Bounds in ℝ≥0 #25 Bounds in ℝ≥0 (convention)
  2. Couplings of random systems #26 Couplings of random systems (systems)
  3. The H-coefficient bound for random systems #27 The H-coefficient bound for random systems (systems)
  4. Games equivalent as games and the fundamental lemma #28 Games equivalent as games and the fundamental lemma (games) ← this PR
  5. Identical until bad: a common part, per-step costs, and coupled games #29 Identical until bad (games)

🤖 Generated with Claude Code

https://claude.ai/code/session_01E7GPy1oUx164NSZsxJT3N9

`RandomSystem.GameEquivalent`, two games equivalent as games, G ≡g H (CR18, Definition 4.16,
p. 105): they give every transcript with every MBO unset the same probability, the cumulative
condition CR18 states as equivalent to identical pre-winning behavior; Maurer 2002,
Definition 5 (F^A ≡ G^C), is the same notion with a monotone event sequence in place of the
unset MBO. Such games are won alike (`GameEquivalent.winProbability_eq`,
`Domain.Distinguisher.winProbability_eq_of_gameEquivalent`; CR18, Lemma 4.15; Maurer 2002,
Lemma 7 and Lemma 6 (ii)), and the advantage between their visible systems is at most the
winning probability, the fundamental lemma (`GameEquivalent.abs_decisionProbability_sub_le`,
`Domain.Distinguisher.advantage_le_winProbability_of_gameEquivalent`; CR18, Lemma 4.16;
Maurer 2002, Lemma 7 and Theorem 1 (i)).

`Distribution.abs_mass_sub_le_of_le`: two laws of weight one above a common part `ν` differ on
every event by at most `1 - ν.weight`. The conditional-equivalence bound
`RandomSystem.abs_decisionProbability_sub_le_winProbability` now uses it instead of proving the
step inline. THEORY.md §1.5 gets a paragraph; `#print axioms` lines for the four theorems.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E7GPy1oUx164NSZsxJT3N9
@MarcIlunga
MarcIlunga changed the base branch from h-coefficient to main October 13, 2026 09:01
@MarcIlunga
MarcIlunga merged commit b6b2049 into main Oct 13, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant