Audit infrastructure

AsPredicted receipt manifest

For external methodology changes filed after 2026-04-25, Convexly attempts to lock hypotheses before analysis and then publishes the receipt status here. External links are shown as verified only when the linked page contains the expected AsPredicted ID, title, and filing date; otherwise they are marked pending, stale, or broken. The verdict + run date + effect-size CI are reported within 24 hours of the test running. The original V1 and V1-M papers were not retroactively pre-registered; they remain frozen-coefficient methodology with ex-ante version-controlled commitment via the SHA-256 audit chain. Failed methodology tests land in the negative-result registry; the audit chain is verifiable in your browser at /research/verify.

Machine-readable manifest at /research/preregistrations.json.

Last updated 2026-05-10T00:49:34Z.Receipts checked 2026-05-10T22:09:46Z.8 entries

Receipt health: 1 externally verified/7 pending public URL/0 broken or stale

AsPredicted #287368

Filed 2026-04-25 · Ran 2026-04-27

V1.5 follow-up experiments E2 + E7

Failed (rejected)
External receipt verified

E2 per-wallet temporal holdout: ρ = +0.111 [+0.046, +0.175], well below the +0.30 pre-reg threshold. E7 per-quarter IC stability: median ρ = +0.038, only 3 of 5 quarters positive vs ≥5/6 required. Both failed.

AsPredicted receipt #287368Edge Score Methodology V1.5: deferred-ex...Registry entry

Audit-chain anchor: v1_5_analyses_results_20260427_191717

AsPredicted #287436

Filed 2026-04-26 · Ran 2026-04-26

MarketAlpha V2 in-sample skill-weighted aggregation tests

Failed (rejected)
Public receipt not verified

Initial in-sample test of skill-weighted aggregation as per-market price prior. 24 aggregator variants tested; all rejected. Cohort substitution amendment filed as #287714.

AsPredicted #287436: public receipt not yet verifiedMarketAlpha V2.8.2: skill-weighted aggre...

Audit-chain anchor: marketalpha_v2_in_sample_run

AsPredicted #287442

Filed 2026-04-26 · Ran 2026-04-26

MarketAlpha V2 forward-only skill-weighted aggregation tests

Failed (rejected)
Public receipt not verified

Forward-only complement to #287436. Same 24 aggregator variants on a held-out forward window. All variants rejected forward.

AsPredicted #287442: public receipt not yet verifiedMarketAlpha V2.8.2: skill-weighted aggre...

Audit-chain anchor: marketalpha_v2_forward_run

AsPredicted #287714

Filed 2026-04-27 · Ran 2026-04-27

MarketAlpha V2 cohort-substitution amendment

Failed (rejected)
Public receipt not verified

Cohort substitution from V1 (8,656 wallets) to V1-M (8,778 wallets) to verify the negative result is not cohort-specific. All 24 aggregator variants rejected on V1-M as well; consistent with the original finding.

AsPredicted #287714: public receipt not yet verifiedMarketAlpha V2.8.2: skill-weighted aggre...

Audit-chain anchor: marketalpha_v2_cohort_substitution

AsPredicted #287983

Filed 2026-04-28 · Ran 2026-04-29

V2.8.2 wash-filter TOST equivalence test on V1-M Polymarket cohort (Sirolly-adapted)

Passed
Public receipt not verified

Wash-filter robustness check on the V2.8.2 negative result. Brier delta CI [+0.16028, +0.19287] sits inside the pre-registered TOST equivalence range [+0.154, +0.204]. Movement after wash filtering: +0.00243 Brier (1.4% relative). The V2.8.2 finding (skill-weighted aggregation rejected) is robust to wash-trader filtering at composite-z >= 3.0.

AsPredicted #287983: public receipt not yet verifiedMarketAlpha V2.8.2: skill-weighted aggre...Registry entry

Audit-chain anchor: v28_2_wash_filter_tost_passed_2026_04_29

AsPredicted #288046

Filed 2026-04-29 · Runs 2026-07-29

CME V0.2 backtest: 90-day walk-forward on Polymarket constraint-projection signals

Pending
Public receipt not verified

90-day walk-forward backtest of the CME V0.2 constraint-projection pipeline pre-registered. Hyperparameters frozen ex-ante (thresholds, sizing, cost model, performance metrics). No hyperparameter tuning based on backtest results allowed by the pre-reg.

AsPredicted #288046: public receipt not yet verifiedCoherent Markets Engine V0.1 + V0.2.0

Audit-chain anchor: cme_v0_2_0_methodology_frozen_commit_8616a63

AsPredicted #288610

Filed 2026-05-01

V2-Perps Edge Score: skill ranking with CRPS + funding-capture pillars

Pending
Public receipt not verified

Pre-registers the form (4 pillars: CRPS-posture, conviction, discipline, funding-capture) + 7 validation gates for the V2-Perps Edge Score composite. Form locked at freeze commit 8c86dd4; coefficients TBD pending Hyperliquid 90-day cohort fit. Composite reduces to V1 / V3b on binary outcomes (Brier-equivalence identity) and extends across crypto perps, equity perps, compute futures, AI benchmark markets, valuation futures, and prediction markets per spec Section 6.

AsPredicted #288610: public receipt not yet verifiedEdge Score V2-Perps: cross-substrate ski...

Audit-chain anchor: v2_perps_methodology_frozen_commit_8c86dd4

AsPredicted #288615

Filed 2026-05-01 · Runs 2026-07-30

CME V0.2-Perps: 90d walk-forward on Hyperliquid coherence-violation signals

Pending
Public receipt not verified

90-day walk-forward backtest of the V0.2-Perps coherence-violation engine (7 constraints: cash-and-carry, triangle, put-call parity, Carr-Madan butterfly, Litterman-Scheinkman PCA calendar, cross-venue 4-corner, vertical-spread monotonicity). H1 net Sharpe > 1.0; H2 capacity ceiling < 50K USD/day; H3 each constraint contributes positive Sharpe with 95% bootstrap CI excluding zero. Methodology code freezes at commit adb99d6; emit cron at .github/workflows/cme-v0-2-perps-emit.yml.

AsPredicted #288615: public receipt not yet verifiedCME V0.2-Perps: cross-venue coherence-vi...

Audit-chain anchor: cme_v0_2_perps_methodology_frozen_commit_adb99d6

Filing policy

  • Filing rule: For post-2026-04-25 methodology changes that affect external claims, Convexly either files a pre-registration before analysis runs or marks the item internal-only / pending-public-url until an external receipt can be verified.
  • Verdict update rule: When a pre-registered test runs, the verdict + run_at_utc + verdict_summary are updated within 24 hours of the run completing. Verdicts are PASSED, FAILED, or PENDING. Failed pre-registrations are added to the negative-result registry at /research/negative-results.
  • Supersession rule: When a pre-registration is superseded by an amendment, the original entry is kept (verdict noted as superseded) and the amendment is added as a separate entry. Original entries are never removed.
  • Audit-chain link: Every entry's audit_chain_anchor field references the SHA-256-hash-chained run identifier in apps/web/public/research/cme/audit_log.jsonl (or paper-specific provenance log). The /research/verify page walks the chain in client-side JavaScript and renders a green stamp if every prev_hash matches its parent's row_hash.