Every published OOB proof state includes the exact game rules, remaining shoe, action EVs and methodology needed to independently check the result.
This hub brings together the documented methodology, internal controls, reproducible states and simulated comparisons. Independent third-party validation has not been performed. No real-play results are published.
Each proof asset answers a different question — from a single reproducible decision to large-scale controlled comparisons.
The checks and reruns below are documented in the Validation Register. Each record states its scope, the build or version it covers, its result and its limits.
| Check | Scope | Version | Result |
|---|---|---|---|
| V-001Agreement with the source spreadsheetInternal control | Six full-shoe 8-deck rule configurations; the integrity group of the manual plan; five composition scenarios. | Engine ported from the source spreadsheet; recorded before build identifiers were introduced. | Passed · Partial |
| V-002Rule restrictions on split handsInternal control | Double after split, double restriction and split-ace rules on every hand created by a split. | Application release of 23 Aug 2026. | Passed · Complete |
| V-003Ten published reproducible states — rerunReproducibility | OOB-RS-0001 to OOB-RS-0010: rebuilt composition, available actions, action EVs to four decimal places, highest-EV action. | Build OOB-ENG-b0681ede84a3; Methodology v1.0. | Match · Complete |
| V-004Build and runtime identificationReproducibility | Build identifier from the engine source files; runtime identifier from the code held in memory. | Build OOB-ENG-b0681ede84a3. | Passed · Complete |
| V-005Comparison with an internal reference programInternal control | Selected states on very small shoes and on shoes of about 290 cards: Stand, Hit, Double, ENHC dealer blackjack, and one split state under the production conventions. | Reference program version a79594ccbc59; engine as of 13 Sep 2026. | Measured · Partial |
| V-006Published house-edge values (Wizard of Odds)Internal control | Three rule points used during calibration; full-shoe values for 8, 6 and 4 decks. | Calibration period, before 6 Sep 2026. | Recorded · Historical |
| V-007Basic Strategy comparison, decision levelInternal control | 813 initial two-card, non-pair states under the default rules (1,000 generated; 131 pairs and 56 naturals excluded by design). | Run bs_vs_oob_full_20260909_182319, seed 20260908. | Measured · Partial |
| V-008Basic Strategy labels of published statesInternal control | The Basic Strategy labels shown for OOB-RS-0003, 0004 and 0005. | Candidate reference chart read on 6 Sep 2026. | Match · Complete |
| V-009Audit of the production pre-bet EVInternal control | Design: 113 rounds drawn from an earlier 300,000-card simulated track. | Audit protocol v1.1. | No verdict · Incomplete |
| V-010First-million accounting reconciliationInternal control | Accounts V1 and K1, every wagered round of the first one-million-card track (seed 20261001). | Simulator OOB-K1-b564264c55fd. | Passed · Complete |
The OOB methodology explains how exact shoe composition, action outcomes and Expected Value are calculated — while implementation-specific optimizations remain proprietary.
Exact composition does not predict the next card. It evaluates the probabilities and Expected Value of the current state.
OOB proof assets move through defined stages before they are presented as validated evidence. Status labels below reflect the current publication stage of each asset.
| Proof asset | What it demonstrates | Status |
|---|---|---|
| Reproducible State Library | Published state-level decisions are designed to be independently inspected and reproduced. | Internally reproduced |
| Same True Count / Different Shoe | Equal Hi-Lo True Counts can correspond to different exact compositions, EVs and sometimes optimal decisions. | Internally reproduced |
| Basic Strategy vs OOB | Static Basic Strategy and composition-dependent OOB decisions are compared under identical game states. | Internally validated |
| One Million Cards | The realised results of OOB Graduated, OOB Selective and Basic Strategy — Flat Bet over two simulated tracks of 1,000,001 cards from the same shuffles, with a fixed $10 unit. | Retrospective analysis |
| One Million Cards: First Track (Archive) | The realised result of the MOV fixed-unit bet ramp, applied retrospectively to one simulated track of 1,000,005 cards with OOB decisions. | Retrospective analysis |
| Technical Whitepaper | The model, its conventions and approximations, and the scope of each piece of evidence are documented, including what the evidence does not establish. | Published document |
| Validation Register | Internal checks and reruns, each with its scope, version and result, and every reported discrepancy and correction. These are internal records, not independent validation. | Published document |
| 100-Hour Simulated Replay | A retrospective comparison of OOB Graduated, OOB Selective and Basic Strategy — Flat Bet over 6,000 simulated rounds, equivalent to 100 hours under the assumption of 60 seconds per round. | Retrospective analysis |
Start with the proof as it is published, then use OOB to see how the exact remaining shoe updates action EVs, the highest-EV decision, betting guidance and risk context in real time.