Objective
Independently replicate the distinct RecessionAlert row About the new TRENDEX SP-500 model | RecessionAlert from the checked-in canonical snapshot, without a live-site request, credentials, private account data, or mutation of data/recessionalert.sqlite.
Approach
The canonical input is the manifest-selected page body in read-only data/recessionalert.sqlite, cross-checked against research/findings/recessionalert_inventory.md and research/artifacts/recessionalert/manifest.json. The page SHA256 is sha256:0847fa3d8bd7a9dc74a9e07e000ec338362ad84c743b57d710fab8f5d4b8119c; it is 80,141 bytes and was fetched at 2026-08-22T23:46:43Z. The article is dated 2019-09-04T10:50:10-0400. The 2544.372604166667-day article-to-snapshot interval is archival delay, not TRENDEX publication or release lag. The crawl log preserves two same-URL page-body vintages; the later digest is manifest-selected, and the target page has zero asset rows.
Disclosed rule and interpretation
The article describes TRENDEX as superseding Demark and Demark+ trend counting. It focuses on support and resistance levels rather than closes four days ago and moving averages, and dynamically maps trend-duration counts to probabilities represented by the historical record of successful trend reversals. It says the method can issue probabilities of market tops and market bottoms.
Its chart legend defines STOP as the cancel level for the current trade, p(n) as the probability of the last day depicted on the chart, and p(n+1) as tomorrow's probability if price does not cross the black line. It says SP500 TRENDEX is a component of the multi-factor Market TOP probability model in the SP500 TOP tab. These are source statements, not an executable action rule. The page gives no support/resistance construction, price field, trend-count initialization, probability calibration, range, threshold, crossing tolerance, persistence, missingness rule, or state transition. It does not disclose a buy-only, sell-only, paired, risk-on, risk-off, or recession-warning mapping.
Native span, publication lag, and causal assumptions
The native evidence is one dated narrative HTML article. No source-linked TRENDEX component or composite series, probability history, or target-page data asset is present, so effective numeric span is not computable. The p(n+1) wording is a next-day conditional forecast horizon, not a release timestamp.
The source does not disclose model calculation time, recurring publication clock, timezone, information cutoff, vintage, or revision schedule. The article-to-crawl interval is not release lag. Unknown dates and values remain unavailable rather than zero, neutral, false, stale-valid, or imputed. Any future overlay requires documented point-in-time input availability and at least one available-bar execution lag; same-close/lookahead is unsupported. Signal assessment remains separate from strategy-overlay or market-performance claims.
Validation and result
The focused offline suite tests/test_recessionalert_trendex_sp500.py passes 8 tests. It verifies manifest/inventory/SQLite digest agreement, selected crawl history, disclosed TRENDEX rule and probability labels, causal lag policy, zero target-page assets, read-only database access, non-playable output, public boundary, and explicit no-series metrics. The executed notebook is research/experiments/1349.ipynb; its offline HTML is generated at docs/experiments/1349-analysis.html by the canonical notebooks producer.
No numeric baseline, extra-lag, subperiod/regime, adjacent-parameter, false-signal, alternative-data, market-outcome, or market-overlay metric is run. The snapshot has no source-linked observations or executable state rule. Demark/Demark+, the separate SP500 Probability Model, the SP500 TOP composite, chart/image OCR, adjacent workbooks, public proxies, synthetic history, or an inferred formula would change or merge the distinct source identity.
Decision: insufficient_evidence. Publish the source-backed concept, provenance, article span, publication-lag status, causal policy, and explicit gaps only. Do not register a playable Other-indicators signal, emit signal history or an equity curve, or claim market validation.
Explicit gaps
No source-linked machine-readable TRENDEX observations, support/resistance values, trend states/counts, probability history, or native numeric span; no exact support/resistance, trend-count, reversal-probability, or state-transition formula; no lookback, smoothing, score/probability range, threshold, black-line crossing rule, or STOP construction; no recurring point-in-time release metadata, timezone, vintage, or revision policy; no buy/sell/hold or top/bottom action mapping, target instrument, allocation, rebalance, holding period, costs, slippage, exit, or re-entry; no independent S&P 500 outcome window; and no captured PRO chart/image bytes or source-linked redistributable data file.
Web plan and boundary
Web plan:
- canonical section: Other / Research via the experiments index;
- discoverable path: generated experiment and analysis links are built by
research.reportsand linked from existing research pages; - browser evidence: the generated evidence table and no-signal/disposition page are the appropriate observable surface;
- static/local boundary: this is an offline static evidence report, not a live TRENDEX evaluator or playable overlay;
- exception: a parameter-control or playable indicator workflow would be misleading because the source supplies no series, thresholds, or action mapping. Reconsider only after a separately authorized refresh provides source-linked observations and executable semantics.
Public output excludes cached HTML, PRO chart/image bytes, uncaptured workbook bytes, credentials, cookies, private account data, and unverified probability or price series.
See research/findings/1349.md, research/findings/specs/recessionalert-trendex-sp500.json, research/experiments/1349.json, and research/experiments/1349.ipynb for the complete evidence record.