Proof, written down.
Permanent notes on the cryptographic verification of track records, for AI agents and the humans beside them: the mathematics we enforce, the failures we publish, and the exact boundary of what a proof can and cannot establish. Published as the work completes, a note per result and a paper roughly every quarter, each at a permanent URL, and nothing ships without a receipt. The standard itself is published: VTR-1 on SSRN, 22 pages, 41 normative requirements, nine conformance vectors, CC BY 4.0.
Written and maintained by Mohammad Muavia, Founder of MIZAN. Each note carries a permanent citation and, where empirical, a worked example you can recompute yourself: the numbers in a note about verification are themselves verifiable.
What “verified track record” means, written as a specification.
The VTR Standard (VTR-1): what “verified track record” means
A formal, versioned, open specification, the way GIPS is for performance reporting. Three conformance levels, normative requirements, the verification procedure, and what it refuses to certify. Anyone may implement it.
Read the full text →One research program, in parts.
Each paper makes one enforceable claim, with its priority timestamped to Bitcoin before release. The capstone, a consolidated reference-implementation paper, comes last, after each part has been public, attacked, and survived.
Who Counts the Trials? A Committed Trial Ledger for Enforcing the Deflated Sharpe Ratio in Zero-Knowledge
The manager’s discretion, removed: N becomes the leaf count of a committed Merkle tree, the winner is forced in-circuit to be the maximum, the deflation recomputes inside the proof. The paper behind everything on this page.
ssrn.com/abstract=7187038 ↗Forty-Six Years of Unverifiable Alpha: The Track-Record Verification Problem in Quantitative Finance — Its Literature, Its Cost, and Its Closure in Zero Knowledge
The whole 46-year literature, assembled and argued in one place, from Grossman–Stiglitz to the modern overfitting program: the price of the gap, why every prior fix failed, and the mechanism we built for it. The claim is now on SSRN.
ssrn.com/abstract=7352018 ↗Who Sees the Strategy? The Sealed Model
Proving a hidden model’s track record without revealing a parameter: committed position streams, forward-only tracks, fail-closed semantics. A mechanism for the disclosure question. Companion note below (№ 06).
Who Judges the Judge? Era Law: Verifier Governance for Credentials That Outlive Their Issuer
The issuer’s discretion, removed: eras named by their hash, superseded never revoked, one byte is a new judge, weakness disclosed rather than rewritten, with the production scars documented and the extension on-chain.
Start here, then check our arithmetic.
Proving the Deflated Sharpe Ratio in Zero-Knowledge
Bailey & López de Prado gave the discipline a way to tell skill from luck: deflate a Sharpe by how many strategies you tried. But a self-reported deflation can be gamed; the person computing it chooses the trial count. We compute the Deflated Sharpe inside a STARK, on committed data, with the trial ledger cryptographically bound, so the correction cannot be gamed by the person it judges. With a worked example you can recompute yourself, and the boundary of the claim.
Read the note → Tool · Reproduce · offline · no MIZAN binaryReproduce it yourself
Don’t trust our verifier. Download ~200 lines of dependency-free Python and the pinned data, run it offline on your own machine, and watch a live credential reproduce to the digit. Both halves are checkable: the arithmetic (recompute the flagship) and the overfitting test (exact CSCV over all 12,870 splits).
Open the reproduction kit →Every note, tool and reference, at a permanent URL.
-
№ 02Method
Does the overfitting statistic have teeth?
A both-sides demonstration: the Probability of Backtest Overfitting must catch a known-lucky search and clear a known-skilled one, or it is theatre. Exact CSCV over all 12,870 splits, on scenarios where the truth is known by construction. Pure noise lands at a coin-flip 0.442; genuine skill collapses to 0.001; our live credential reads 0.0759.
PBOCSCVRead the note → -
№ 03Boundary
What we refuse to certify
The exact line between what a MIZAN credential proves and what it does not: historical seal versus anchored forward track, a coverage map of the strategy universe (thirteen classes provable today, further classes announced only when they ship, one refused on principle), and why a credential is only worth what its issuer refuses to sign.
CoverageRead the note → -
№ 04Discipline
We publish the strategies that failed
A verification layer is only credible if it publishes its losers as readily as its winners. MIZAN mints real cryptographic proofs of strategies that failed the gate, the first GPU mint among them, and anchors them beside the passes. Why the honest failure is the most important credential we issue.
Published FAILRead the note → -
№ 05Contributions
What’s new here, and what isn’t
MIZAN’s technical contributions, written as claims you can check rather than adjectives you have to trust: the anti-overfitting framework enforced in zero-knowledge, a committed trial ledger that answers “who counts the trials,” both schools of historical-test honesty proven in-circuit, era-pinned and self-failing credentials, and, stated with equal care, what we did not invent.
What's newRead the note → -
№ 06Privacy
Prove it without revealing it
The other half of the problem. Proving a strategy is honest is worth nothing if proving it means showing it. The sealed model lets a quant run any strategy (ML, code, discretion) anywhere, submit only its bar-by-bar decisions, and prove the performance while the logic never leaves their machine. With the honest limits of what a seal does and does not establish.
Sealed modelRead the note → -
№ 07Receipt
The gate refused its own maker
We ported our own trend engine to the S&P 500, ran it through our own locked gate on survivorship-free US equity data, and the gate said no, twice: first refusing our mis-declared annualization basis, then the strategy itself. The full cryptographic receipt, downloadable and timestamped to Bitcoin, and why a refusal is the strongest credential an issuer can publish about itself.
Published FAILRead the note → -
№ 08Instrument
Reading a verdict
Seventeen checks, 81 milliseconds, one word. A line-by-line read of the live PASS credential by its own issuer: the Merkle-pinned data, the locked gate, the sealed strategy, the out-of-sample decay shown rather than buried, and, at equal length, what a verdict does not tell you.
VerificationRead the note → -
№ 09Semantics
What a PASS actually asserts
The most dangerous thing about a credential is not forgery; it is over-reading. The issuer bounds his own first gate-PASS with legal care: the complete in-circuit assertion set of BTC-TREND-4H, and, at equal length, exactly what it does not assert: future returns, capacity, fills, transfer, or deflation for search multiplicity.
VerificationRead the note → -
№ 10Lineage
Twenty authors, one machine
The complete citation ledger of the engine: Sharpe’s ratio, the Bailey–López de Prado deflation program, the White–Hansen–Harvey–Liu multiple-testing program, the forking-paths critique, and the cryptography, with exact references and what the circuit enforces from each. Attribution, never endorsement.
FoundationsCitationsRead the note → -
Free toolCalculator
Deflated Sharpe Ratio calculator
Is your Sharpe real, or the luckiest of N tries? Compute PSR, DSR and minimum backtest length (the Bailey–López de Prado corrections) live in your browser, nothing uploaded. With the honest disclosure every such calculator owes you: the correction is only as honest as the self-reported N.
CalculatorOpen the calculator → -
ReferenceQuestions
Can you prove a backtest without revealing the strategy?
Straight answers to what quants, allocators, and researchers actually ask: how to prove a historical record is real without revealing it, enforce the Deflated Sharpe cryptographically, verify a track record trusting no one, prove an AI model’s returns without exposing it. Every answer links to a proof you can re-verify yourself.
QuestionsRead the answers → -
ForwardableAllocators
Seven questions an allocator should ask any manager
Seven questions to put to any manager pitching any track record, in any asset class, none of which depend on MIZAN existing. Why each matters, what an honest answer sounds like, and the tell when someone retreats into vocabulary.
QuestionsAllocatorsRead the questions →
Nothing in this view.
Cite the record. Trace what it stands on.
- Muavia, M. (2026). VTR-1: The Verifiable Track Record Standard: Enforcing Backtest Honesty Without Trusting the Issuer or Revealing the Strategy. SSRN. ssrn.com/abstract=7215664
- Muavia, M. (2026). Who Counts the Trials? A Committed Trial Ledger for Enforcing the Deflated Sharpe Ratio in Zero-Knowledge. SSRN. ssrn.com/abstract=7187038
- Muavia, M. (2026). Forty-Six Years of Unverifiable Alpha: The Track-Record Verification Problem in Quantitative Finance. SSRN. ssrn.com/abstract=7352018
Twenty authors, one machine.
Nothing here starts from zero. The lineage note is the complete citation ledger of the engine, from Sharpe’s ratio through the Bailey–López de Prado and White–Hansen–Harvey–Liu programs to the cryptography, with what the circuit enforces from each. Attribution, never endorsement: none of the authors cited has reviewed or endorsed MIZAN.
Muavia, M. (2026) · SSRN