Browse › Probability and statistical mechanics › 238 Optimal logarithmic mixing of the Thorp shuffle
Optimal-order mixing of the Thorp shuffle
Open PDFCite Report a problem Add a reference or comment Interested in working on this Start a discussion
Edit NOTES.md · Existing issues · Source on GitHub
Abstract
We prove that the Thorp shuffle on $2^d$ cards mixes in $\Theta(d)$ complete shuffles. The full permutation law after $1600d$ shuffles converges to uniform in total variation as $d\to\infty$, uniformly over the initial deck, while a support count gives a lower bound of $2d-O(1)$.
If the PDF does not display here, open it directly: Open PDF.
Notes
No notes yet. Add a reference, a comment or your name with the buttons above, or edit NOTES.md on GitHub.
Automatically collected notes (unreviewed)
work in progress
Collected automatically by AI tools (Claude Opus 5.5 and GPT-6 Astra, 2026-10-21) and not checked by a person. May be wrong or incomplete. Report an error.
Follow-up works
- acprk (GitHub), Remark R5: Thorp shuffle 1600d → 24d (openai-math-remarks, local draft, undated). Re-selects parameters in this paper's first stage and combines them with the companion's Fourier transfer to claim mixing after 24d shuffles (45d using this paper alone); asymptotic, not formalized, reviewed only by AI. [AI-generated per source]
- public-observatory (GitHub), thorp-shuffle-cutoff: an open research agenda on cutoff for the Thorp shuffle (GitHub repository, October 2026). A research agenda on whether the Thorp shuffle has cutoff at 2·log₂N, taking Theorem 1.1 of this paper as its starting point; its reading notes (2026-10-11) say the paper leaves the leading constant and cutoff open. [AI-generated per source]
Formalization
- OpenAI math release, Lean scope notes for family 238 (lean/docs/238.md, 2026). States that the formalization proves full-deck total-variation convergence after $1600d$ shuffles together with the 2d−O(1) support lower bound, using the comparator files ThorpFirstReciprocal.lean and ThorpRemaining.lean. [family-level]
Possibly relevant literature (found by the agent, not checked)
- Y. Jiang, A subquadratic upper bound for the mixing time of the Thorp shuffle (arXiv:2610.18884, 3 Oct 2026). Independent preprint proving a full-deck bound of order d^{5/3} for n=2^d, improving Morris's O(d^3); relevant for comparison with Theorem 1.1's O(d) bound. It does not mention the release. [family-level]
Digests and expositions
- Latent Seminar, Optimal logarithmic mixing of the Thorp shuffle (expository notes, 2026-10-06). Outlines all twelve manuscripts’ proof methods and formalization boundaries, explicitly disclaiming independent full-proof verification. [family-level]
- heyprof, Thorp shuffle: how often until the deck is random? (interactive exposition, undated). Illustrates the family’s full-deck mixing question with small-deck distributions, single-card marginals, and the support bound. [family-level]
News and blogs
- Joe Halliwell and Claude Opus 5.5, Does any of the OpenAI maths have real-world consequences? (blog, 2026-10-08). Discusses this paper’s $1600d$ bound in connection with format-preserving encryption and its limited immediate practical implications. [AI-generated per source]
Elsewhere
- VibeMathed: Problem record
- Emergent Mind: family 238
Independent sites; their content is not reviewed here.
Cite
@misc{OAI:Optimal-order-mixing-of-the-Thorp-shuffle-September-26-2026,
author = {{OpenAI}},
title = {{Optimal-order mixing of the Thorp shuffle}},
howpublished = {OpenAI Math Release preprint
\href{https://github.com/openai/math/blob/main/preprints/Optimal-order-mixing-of-the-Thorp-shuffle-September-26-2026/paper.pdf}{OAI:Optimal-order-mixing-of-the-Thorp-shuffle-September-26-2026}},
year = {2026}
}Lean formalization
OpenAI provides Lean files for this family: lean/docs/238.md. The Lean statements have not been checked by people against the paper's claims.
Family 238: Optimal logarithmic mixing of the Thorp shuffle
Proves that the Thorp shuffle randomizes $N=2^d$ labeled cards in $\Theta(\log N)$ physical shuffles, settling its optimal mixing order for power-of-two deck sizes. Convergence is in total variation from the worst initial ordering and concerns the entire permutation, not just individual card positions.
Other papers in this family:
- Random coordinate frames and partial permutation laws · September 26, 2026
- From partial permutation information to Fourier bounds · September 26, 2026
- Conditional information under deterministic coordinate sweeps · September 26, 2026
- Conditional permutations in a revealed switching environment · September 26, 2026
- Routing densities and representation contraction for Thorp sweeps · September 26, 2026
- Row–column symmetry and contraction of coordinate sweeps · September 26, 2026
- Random-subspace tests and trace smoothing for coordinate sweeps · September 26, 2026
- Compatibility entropy and the spectrum of a Thorp sweep · September 26, 2026
- Signed tensor densities and diagram budgets for the Thorp shuffle · September 26, 2026
- Conditional coordinate sweeps and analytic transfer · September 26, 2026
- A strict four-row permanent inequality and permutation moments · September 26, 2026