Every project, its stage, and the evidence for it.
Stages follow the Lab status vocabulary, in order of how much can be claimed. A project moves up only when evidence is attached, may stop at any stage, and is recorded as failed rather than removed when it does not survive validation.
| Project | Stage | Evidence | State as of |
|---|---|---|---|
| Validating — results are being checked against the pre-registered question | |||
| Keller trend model | Validating | Independent replication of David Keller’s three-timeframe weekly trend model; not affiliated · runs on the S&P 500 index daily close (FRED) · 44 tests · pass test set before the first check: 8 consecutive weeks matching StockCharts’ chart · 1 of 8 weeks matched · 4 of 6 historical checks match Keller’s published calls, 2 discrepancies under investigation · readings unpublished until validation passes | 2026-09-26 |
| Testing — implementation exists and is being run against data | |||
| Momos rank | Testing | Matched the published StockCharts SCTR at a rank correlation of 0.9953 across 483 names on 2026-07-02; no pass threshold was set in advance · 5.5M panel rows · 201 tests · runs 18:00 each weekday, board published after it · latest board 2026-09-25 | 2026-09-26 |
| Momos hook pool | Testing | +2.3% over 20 days out-of-sample against +0.8% for a control, 856 cases in the S&P 1500 (2026-07-10 article) · concentrated in the last twelve months · forward record builds each weekday · no pre-registration of entry rules yet | 2026-09-26 |
| Ratio board | Testing | 81 symbols and ratios, 80 scored · run weekly on the Friday close, published by hand · latest board week ending 2026-10-02 · no pre-registered question yet · scorer not yet under version control, no automated tests | 2026-09-26 |
| Concept — the question exists. No scope, no method. | |||
| Cycle dashboard | Concept | Display built on the main site with sample figures · publishing gated on verified cycle anchor inputs · no validation approach written yet | 2026-09-26 |
| Failed — did not survive validation; recorded, not removed | |||
| Momos trajectory ranking as a return predictor | Failed | Pre-registered test: top decile minus bottom decile straddles zero, in-sample and out-of-sample, across every group and horizon (2026-07-10 article). Kept as a screen, not an edge. | 2026-07-10 |
Stage definitions
Concept
The question exists. No scope, no method.
Specced
Purpose, boundaries, and validation approach are written down. No indicators selected, nothing built.
Researching
Inputs are being selected and justified against the spec.
Testing
Implementation exists and is being run against data.
Validating
Results are being checked against the pre-registered question, including for a null result.
Candidate
Validated. Under evaluation for production use.
Graduated
Moved to the primary King Nuggets research product.
Failed
A project that fails validation is recorded as failed, not removed.
Definitions are quoted from LAB_STATUS.md. Stage is carried by the label and the grouping. It is never encoded in colour.