{
  "schemaVersion": "1.6",
  "slug": "signed-permutation-prefix-balancing",
  "title": "Prefix balancing with signs and coordinate permutations: the exact constant 5/2 in dimension four and online bounds above 5/2",
  "shortTitle": "Flip and shuffle: the balancing constant is 5/2 in four dimensions",
  "url": "https://evidencepress.org/releases/signed-permutation-prefix-balancing/",
  "oneLine": "Five fixed vectors force a running total to reach 5/2 in every even dimension from four, so when each vector may be sign-flipped and shuffled the best balancing constant in dimension four is exactly 5/2; online, 5/2 cannot be beaten from dimension four and is exceeded from dimension eight.",
  "abstract": "Lists (vectors) of n numbers in [−1, 1] arrive one at a time and are added to a running total. Before adding each one you may flip all its signs and permute its coordinates, but not reorder the stream; the aim is to keep the largest absolute coordinate of every running total small. With signs alone the best guarantee grows like √n. Niles-Weed, Sadovsky and Shkrob (arXiv:2610.02127, 2026) showed that with permutations 3 − 2/n always suffices, even choosing on the fly, recorded that 2 cannot be beaten, and asked for the sharp constant. This candidate shows that five fixed vectors force a running total of size at least 5/2 in every even dimension n ≥ 4; in dimension four this meets the upper bound, so the sharp constant there is exactly 5/2, offline and online. For deterministic online strategies the constant is at least 5/2 in every dimension n ≥ 4, strictly above 5/2 in every dimension n ≥ 8 (for example at least 5/2 + 1/4096 when n = 8) and at least 161/64 = 2.515625 for every n ≥ 52, via a plateau amplification lemma: two coordinates tied at the maximum can be pushed higher by a short geometric sequence of vectors. A computer-checked certificate gives 13/6 online in dimension three. Limits: whether 5/2 can be beaten offline, and whether 3 − 2/n is sharp in every dimension, remain open; the 5/2 obstruction concerns running (prefix) totals, not only the final total; margins above 5/2 are tiny except the 161/64 bound. AI-assisted, unrefereed candidate; no external peer review, independent reproduction or formal verification is claimed.",
  "datePublished": "2026-11-15",
  "dateModified": "2026-11-15",
  "version": "0.1.0-candidate",
  "doi": "10.5281/zenodo.24239377",
  "doiUrl": "https://doi.org/10.5281/zenodo.24239377",
  "conceptDoi": "10.5281/zenodo.24239376",
  "pdfUrl": "https://github.com/ipitchford/signed-permutation-prefix-balancing/releases/download/v0.1.0-candidate/signed-permutation-prefix-balancing-0.1.0-candidate.pdf",
  "altPdfUrl": null,
  "altPdfLabel": null,
  "zenodoUrl": "https://zenodo.org/records/24239377",
  "repoUrl": "https://github.com/ipitchford/signed-permutation-prefix-balancing",
  "releaseUrl": "https://github.com/ipitchford/signed-permutation-prefix-balancing/releases/tag/v0.1.0-candidate",
  "markdownUrl": "https://evidencepress.org/releases/signed-permutation-prefix-balancing/index.md",
  "bibtexUrl": "https://evidencepress.org/releases/signed-permutation-prefix-balancing/cite.bib",
  "audioUrl": "https://evidencepress.org/assets/audio/signed-permutation-prefix-balancing.mp3",
  "imageUrl": "https://evidencepress.org/assets/og/signed-permutation-prefix-balancing.png?v=b0c43c5d88",
  "coverArtUrl": "https://evidencepress.org/assets/art/signed-permutation-prefix-balancing.svg?v=939f683ae6",
  "media": [
    {
      "type": "audio",
      "url": "https://evidencepress.org/assets/audio/signed-permutation-prefix-balancing.mp3",
      "name": "Plain-English audio briefing",
      "description": "Communication aid; not additional research evidence.",
      "transcriptUrl": "https://evidencepress.org/assets/audio/signed-permutation-prefix-balancing.txt"
    }
  ],
  "authors": [
    "Anonymous"
  ],
  "license": "CC0-1.0",
  "status": "unrefereed-candidate",
  "verification": {
    "peerReviewed": false,
    "independentlyReproduced": false,
    "formallyVerified": false,
    "internallyReplayed": true,
    "detail": "Unrefereed AI-assisted mathematics candidate. Theorem 3.1, Corollary 4.3, Proposition 5.2, Lemma 5.3, Theorem 5.4 and Corollary 5.5 have written proofs; Proposition 6.1 (13/6 online in dimension three) is a computer-assisted proof by an exact certificate. Internal AI-assisted review in five reviewer roles and an adversarial review by a second AI model, clean replays, one revision round with three extensions (the online bound for every n ≥ 4, the plateau lemma for every n ≥ 8 and the 161/64 bound, the dimension-three certificate), then a re-review, a proof audit and a second-model audit of the new proofs; two text-level corrections followed, and no numbered result changed. External specialist review, independent reproduction and formal verification have not been established."
  },
  "assurance": [
    {
      "dimension": "availability",
      "label": "Availability and archiving",
      "question": "Is the evidence package publicly retrievable from an archive under a persistent identifier?",
      "state": "passed",
      "evidenceUrl": "https://zenodo.org/records/24239377",
      "note": "The versioned research package and paper are publicly archived on GitHub and Zenodo. Provider asset parity was checked against frozen local files; availability does not establish correctness or peer review."
    },
    {
      "dimension": "internalReplay",
      "label": "Internal replay",
      "question": "Does the producer’s own pipeline reproduce the stated result from the archived package?",
      "state": "passed",
      "note": "verify.py was rerun from fresh copies in normal mode and under python -O; verification.json matched the shipped file byte for byte. A changed expected dimension-four optimum makes the checks fail under -O. GitHub CI reruns verify.py and the negative control (blocking) and explore.py (non-blocking) on Linux."
    },
    {
      "dimension": "independentRerun",
      "label": "Independent rerun",
      "question": "Has someone else run the supplied implementation and obtained the stated result?",
      "state": "not-assessed",
      "note": "No unaffiliated rerun has been established."
    },
    {
      "dimension": "independentReimplementation",
      "label": "Independent reimplementation",
      "question": "Has someone else reached the result from an independent implementation?",
      "state": "not-assessed",
      "note": "No unaffiliated reimplementation; an internal second-model audit reproduced the dimension-three certificate (782 states, zero-state rank 11) and the final-sum optimum 3/2 with its own enumeration."
    },
    {
      "dimension": "formalVerification",
      "label": "Formal verification",
      "question": "Is a formalised statement machine-checked, and over which trusted base?",
      "state": "not-assessed",
      "note": "No proof-assistant formalisation."
    },
    {
      "dimension": "specialistReview",
      "label": "Specialist review",
      "question": "Has a domain specialist assessed the argument?",
      "state": "not-assessed",
      "note": "AI-assisted internal and adversarial review is not external specialist review."
    },
    {
      "dimension": "editorialPeerReview",
      "label": "Editorial peer review",
      "question": "Has a journal or venue run peer review to a decision?",
      "state": "not-assessed",
      "note": "No external editorial peer review."
    },
    {
      "dimension": "dataEnvironmentReproducibility",
      "label": "Data and environment reproducibility",
      "question": "Are data and computational environment pinned well enough to rebuild?",
      "state": "passed",
      "note": "Standard-library Python (3.10 or later); every decision uses exact integers or fractions and the randomised sanity runs use fixed seeds, so the output is platform-independent. Not a locked container."
    }
  ],
  "provenance": {
    "aiGenerated": true,
    "aiAssisted": true,
    "generatedBy": [
      "AI-assisted mathematical research",
      "AI-assisted internal review (five reviewer roles), adversarial review and audit by a second AI model, re-review, proof audit, clean replays",
      "Claude Code: publication packaging and release"
    ],
    "humanRole": "Publication authority only; no human research contribution inferred.",
    "reviewStatus": "Internal AI-assisted and adversarial review, not external peer review."
  },
  "attributionBoundary": null,
  "problem": null,
  "orderedProductAverage": null,
  "corrections": [],
  "publicCorrections": [],
  "publicCorrectionIndexes": null,
  "pageStructureVersion": "2.0",
  "pageStructureVariant": "canonical-theorem",
  "pageStructureWaivers": [],
  "recordMaturity": "current",
  "metadataProvenance": "Prepared from the final paper (revision of 15 November 2026 with the DOI added to the title-page date line, 9 pages), its script and recorded output and the frozen release archive. The versioned GitHub and Zenodo research archives were checked against the frozen release assets on 15 November 2026.",
  "grandfatheredAtSchemaVersion": null,
  "keywords": [
    "vector balancing",
    "discrepancy",
    "prefix discrepancy",
    "online algorithms",
    "signed permutations",
    "Steinitz problem",
    "sharp constant",
    "computer-assisted proof"
  ],
  "keyResults": [
    "Theorem 3.1 (proof): for every even n ≥ 4, five fixed vectors force a running total with a coordinate of absolute value at least 5/2 under every choice of signs and coordinate permutations; in dimension four they are (−1,−1,−1,1), ½(1,1,1,1), (−1,−1,1,1), (1,1,1,1), (−1,−1,1,1).",
    "Corollary 4.3: the sharp constant in dimension four is exactly 5/2, offline and online, matching the upper bound 3 − 2/n of Niles-Weed, Sadovsky and Shkrob.",
    "Proposition 5.2 and Theorem 5.4 (proofs): for deterministic online strategies the constant is at least 5/2 for every n ≥ 4 and at least 5/2 + 2^{1−N(n)} > 5/2 for every n ≥ 8 (for example 5/2 + 1/4096 at n = 8).",
    "Corollary 5.5 (proof): the online constant is at least 161/64 = 2.515625 for every n ≥ 52.",
    "Proposition 6.1 (computer-assisted proof): the online constant in dimension three is at least 13/6 (782-state certificate)."
  ],
  "reviews": [],
  "evidencePackage": "A 9-page paper with a plain introduction, a table of known and new bounds, written proofs of every result except the dimension-three bound, and a section on what remains open; verify.py, a standard-library script of exact checks (about one minute): move-generator cross-checks against brute force, exhaustive enumeration of the five-vector input for even n ≤ 10, every player response to the opening for 8 ≤ n ≤ 16, the full dimension-eight online game end to end, the parameter table and the 782-state dimension-three certificate; and explore.py, bounded exploratory searches reported as negative results. Clean replay from fresh copies in normal and python -O mode reproduced verification.json byte for byte, and a changed expected optimum makes the checks fail.",
  "openProblems": [],
  "mathObjects": [],
  "relatedWorks": [
    {
      "citation": "Niles-Weed, Sadovsky and Shkrob (2026), An optimal constant for vector balancing with permutations. arXiv:2610.02127v1. The source of the problem: the bound 3 − 2/n (Theorem 9), the lower bound 2 (Remark 3) and the question of the sharp constant.",
      "url": "https://doi.org/10.48550/arXiv.2610.02127",
      "doi": "10.48550/arXiv.2610.02127"
    },
    {
      "citation": "Spencer (1985), Six standard deviations suffice. Transactions of the American Mathematical Society. Signing n vectors with error O(√n): the classical balancing scale that permutations remove.",
      "url": "https://doi.org/10.1090/s0002-9947-1985-0784009-0",
      "doi": "10.1090/s0002-9947-1985-0784009-0"
    },
    {
      "citation": "Spencer (1977), Balancing games. Journal of Combinatorial Theory, Series B. The origin of online balancing against an adversary.",
      "url": "https://doi.org/10.1016/0095-8956(77)90057-0",
      "doi": "10.1016/0095-8956(77)90057-0"
    },
    {
      "citation": "Bárány and Grinberg (1981), On some combinatorial questions in finite-dimensional spaces. Linear Algebra and its Applications. Steinitz-type control of running totals.",
      "url": "https://doi.org/10.1016/0024-3795(81)90085-9",
      "doi": "10.1016/0024-3795(81)90085-9"
    },
    {
      "citation": "Banaszczyk (2012), On series of signed vectors and their rearrangements. Random Structures & Algorithms. Prefix control of signed series.",
      "url": "https://doi.org/10.1002/rsa.20373",
      "doi": "10.1002/rsa.20373"
    },
    {
      "citation": "Banaszczyk (1998), Balancing vectors and Gaussian measures of n-dimensional convex bodies. Random Structures and Algorithms. The Gaussian-measure method for the final-sum problem.",
      "url": "https://doi.org/10.1002/(sici)1098-2418(199807)12:4<351::aid-rsa3>3.0.co;2-s",
      "doi": "10.1002/(sici)1098-2418(199807)12:4<351::aid-rsa3>3.0.co;2-s"
    },
    {
      "citation": "Bansal, Jiang, Meka, Singla and Sinha (2022), Prefix Discrepancy, Smoothed Analysis, and Combinatorial Vector Balancing. ITCS 2022, LIPIcs 215. Prefix discrepancy.",
      "url": "https://doi.org/10.4230/LIPIcs.ITCS.2022.13",
      "doi": "10.4230/LIPIcs.ITCS.2022.13"
    },
    {
      "citation": "Ambrus and Bozzai (2024), Colorful vector balancing. Mathematika. The framework in which one vector is chosen from each orbit.",
      "url": "https://doi.org/10.1112/mtk.12274",
      "doi": "10.1112/mtk.12274"
    },
    {
      "citation": "Bansal, Jiang, Singla and Sinha (2019), Online Vector Balancing and Geometric Discrepancy. arXiv:1912.03350. Online discrepancy without permutations.",
      "url": "https://arxiv.org/abs/1912.03350",
      "doi": "10.48550/arXiv.1912.03350"
    },
    {
      "citation": "Alweiss, Liu and Sawhney (2020), Discrepancy Minimization via a Self-Balancing Walk. arXiv:2006.14009. Online discrepancy without permutations.",
      "url": "https://arxiv.org/abs/2006.14009",
      "doi": "10.48550/arXiv.2006.14009"
    },
    {
      "citation": "Anonymous (2026), Linear discrepancy for vector-labelled perfect matchings. Evidence Press. See also only: another discrepancy release, on a different problem (rounding fractional matchings).",
      "url": "https://doi.org/10.5281/zenodo.23932844",
      "doi": "10.5281/zenodo.23932844"
    }
  ],
  "claimAssurance": null,
  "operatingModel": {
    "version": "1.0",
    "workId": "ep-work:signed-permutation-prefix-balancing",
    "attemptIds": [
      "ep-attempt:signed-permutation-prefix-balancing-publication"
    ],
    "aims": [
      "science"
    ],
    "artifactRoles": [
      "research-output",
      "communication"
    ],
    "lineageId": null,
    "accelerationPrimitives": [
      "counterexample-proxy-first",
      "adversarial-controls",
      "assurance-vector",
      "agent-readable-research-object"
    ],
    "decisionObject": {
      "type": "other",
      "description": "Treat the sharp prefix-balancing constant with signs and coordinate permutations as exactly 5/2 in dimension four, at least 5/2 for even n ≥ 4 offline, and above 5/2 online for every n ≥ 8.",
      "scope": "Running (prefix) totals of vectors in [−1, 1]^n under the largest-coordinate norm; the offline constant above 5/2 and the sharpness of 3 − 2/n remain open."
    },
    "bottleneckTargeted": [
      "assurance",
      "publication"
    ],
    "semanticBridge": {
      "state": "explicit",
      "description": "Section 1 and Table 1 separate what Niles-Weed, Sadovsky and Shkrob proved (the problem, the upper bound 3 − 2/n and the lower bound 2) from what is new (the five-vector obstruction, the exact constant in dimension four, the online bounds above 5/2 via the plateau lemma, and the dimension-three certificate), and place the problem in the classical balancing literature.",
      "remainingRisks": [
        "The bounded literature search is not exhaustive priority certification.",
        "The source of the problem is a recent arXiv preprint."
      ]
    },
    "humanJudgmentGates": [
      "Publication is authorised; scientific endorsement is not inferred.",
      "External specialist review and independent reproduction remain open."
    ],
    "parentLinks": [],
    "assuranceTarget": {
      "dimensions": [
        "specialistReview",
        "independentRerun",
        "formalVerification",
        "priorityAssessment"
      ],
      "nextAction": "Have a specialist in discrepancy theory check Theorem 3.1, Lemma 5.1 and the plateau amplification lemma (Lemma 5.3), and rerun verify.py independently.",
      "claimCeiling": "Written proofs of the five-vector obstruction, the exact constant in dimension four and the online lower bounds; computer-assisted proof of the dimension-three bound; internal verification is not external assurance."
    },
    "impactClaims": [
      {
        "id": "signed-permutation-prefix-balancing-no-impact",
        "aim": "science",
        "outcome": "An openly inspectable proof that the sharp prefix-balancing constant with signs and coordinate permutations is 5/2 in dimension four.",
        "setting": "AI-assisted mathematical research",
        "status": "NO_IMPACT_EVIDENCE",
        "designClass": "none",
        "comparator": "None registered.",
        "estimand": "No research-speed or scientific-impact effect is estimated.",
        "evidenceRefs": [],
        "registeredDesignRef": null,
        "independentAssessment": null
      }
    ]
  },
  "researchMetrics": {
    "schemaVersion": "1.0",
    "policyUrl": "https://evidencepress.org/api/research-metrics-policy.json",
    "outcomePolicyUrl": "https://evidencepress.org/api/v2/research-metrics-policy.json",
    "workLedgerUrl": "https://evidencepress.org/api/work-ledger.json",
    "claimCeiling": "These records describe scoped workflow inputs and outcomes. They support calibration and reference-class construction, but do not by themselves establish research acceleration, theorem correctness, novelty, independent assurance or impact.",
    "attempts": [
      {
        "attemptId": "ep-attempt:signed-permutation-prefix-balancing-publication",
        "status": "active",
        "resultClass": "pending",
        "metrics": {
          "schemaVersion": "1.0",
          "measurementScope": "publication-only",
          "scopeBoundary": "Only archive packaging, site integration, guarded delivery and canonical readback from registration are prospective. Earlier research, review, revision and extension work are left-censored.",
          "forecast": {
            "frozenAt": "2026-11-15T05:25:52Z",
            "procedureClass": "full-candidate-publication",
            "targetOutcome": "Accurately scoped GitHub, Zenodo and canonical Evidence Press release with explanatory media and machine readback, if all publication gates pass.",
            "expectedActiveMinutes": 90,
            "plausibleLowMinutes": 60,
            "plausibleHighMinutes": 150,
            "expectedUnattendedWaitMinutes": 30,
            "referenceClass": {
              "label": "Full mathematical candidate publication",
              "sampleSize": 0,
              "basis": "Route benchmark planning prior, not empirical calibration."
            },
            "fermiComponents": [
              {
                "component": "Research archive packaging, replay and identities",
                "count": 1,
                "lowMinutesPerUnit": 25,
                "centralMinutesPerUnit": 35,
                "highMinutesPerUnit": 60,
                "basis": "One frozen package with a rating-free public export, clean replay and DOI insertion."
              },
              {
                "component": "Page, metadata and media integration",
                "count": 1,
                "lowMinutesPerUnit": 20,
                "centralMinutesPerUnit": 30,
                "highMinutesPerUnit": 45,
                "basis": "One reader-first release, drafted in parallel with verification."
              },
              {
                "component": "Guarded deployment and canonical readback",
                "count": 1,
                "lowMinutesPerUnit": 15,
                "centralMinutesPerUnit": 25,
                "highMinutesPerUnit": 45,
                "basis": "Mandatory site tests, one merge cycle with ledger carry-forward, zero-cost deployment and public checks."
              }
            ],
            "probabilityPositiveSignal": 0.9,
            "probabilityTargetClosure": 0.85,
            "probabilityHorizonMinutes": 300,
            "assumptions": [
              "The approved scientific package remains valid and configured providers remain available."
            ],
            "reforecastTriggers": [
              "A publication gate reveals a scientific or integrity defect, or a provider outage blocks delivery."
            ],
            "stopRule": "Do not weaken assurance, integrity or cost gates to meet the forecast."
          },
          "outcome": null
        },
        "missingnessReason": null
      }
    ]
  }
}