Key result: Edge range vs matched control: -1.9pp to +1.3pp.
How it was tested: the record does not state separate training and test periods.
Horizon: 10/20 sessions from the pattern's completion bar. Coverage: ~1,700-name panel (S&P 500 + Nifty + curated watchlists); symbols need >=300 bars.
Stage: Context-only. Version v1; the hub record is as of 6 Nov 2026.
The verdict
NOT RELIABLE as a trading signal — the detector's own rigorous, correctly-clustered audit found 0 of 15 tests survive correction and 2 patterns invert — but the audit methodology itself is a model of how to do this kind of test properly.
Development summary: A rigorous 2026-07-15 audit vs a matched control (prior-move gate, trend x 200-SMA side x regime x region) found NONE of 13 patterns beat it: edges span -1.9pp to +1.3pp, 0 of 15 tests survive correction, 2 patterns invert.
What the review allows
As a descriptive 'what shape printed' read, exactly as the tab's own copy directs.
The date-clustered, matched-control audit methodology as a template for auditing other pattern-style detectors on this dashboard.
What the review rules out
Trade any of the 13 patterns on the edge10 numbers shown — 0 of 15 tests survive correction.
Read Hammer or Three White Soldiers as bullish confirmation — both measure inverted.
Use naive (non-date-clustered) significance on this kind of overlapping-window pattern data — it would falsely pass 6 of 13 patterns here.
What it predicts
Item
Detail
Question it answers
Which of 13 classic 1-3 bar candlestick patterns completed in the last 3 sessions, and does a matched-control audit support treating any of them as an edge?
Method
Candlestick pattern + EB pooling
How it works
Quantitative ATR(14)-unit pattern definitions with a prior-move gate -> forward 10/20-session outcome -> hierarchical empirical-Bayes pooling (pattern x 200-SMA side, own->region->global)
The hub's standard metrics for this model class (only those the record measured).
Metric
Sample / window
Compared with
Value
Range (CI)
Cells surviving
2026-07-15 audit · BH, date-clustered
15 tested
0
n/a
What the validation review found
Is the design sound? Strong: quantitative ATR-unit definitions, a prior-move gate measured strictly before the pattern, a matched control, and date-clustered CIs to correct for overlapping-window autocorrelation — better methodology than most of this dashboard's pattern-style engines.
Backtest. 13 patterns x multiple outcome cuts = 15 tests, all against a matched control, correctly clustered by date; 0 survive. The naive-vs-clustered comparison (6 would falsely pass) is itself a useful, well-evidenced methodological finding.
Limitations
0 of 15 audited tests survive multiple-testing correction; edges span only -1.9pp to +1.3pp against the matched control.
Hammer and Three White Soldiers measure INVERTED (slightly underperform, not confirm, their claimed direction).
Naive, non-clustered testing would have falsely passed 6 of 13 patterns — a real methodological trap this audit specifically avoided.
Fib-zone context carries no measured edge (2026-07-10 audit) and is never attached to a probability.
Questions
Is the Candlestick Pattern Scanner validated?
Its validation review dated 3 Sep 2026 rated it NOT RELIABLE as a trading signal — the detector's own rigorous, correctly-clustered audit found 0 of 15 tests survive correction and 2 patterns invert — but the audit methodology itself is a model of how to do this kind of test properly.
What does the Candlestick Pattern Scanner predict?
It answers: Which of 13 classic 1-3 bar candlestick patterns completed in the last 3 sessions, and does a matched-control audit support treating any of them as an edge? Output: Per-symbol pattern flag (completed within 3 sessions) with edge10 (measured drift in the pattern's own claimed direction). Horizon: 10/20 sessions from the pattern's completion bar. Coverage: ~1,700-name panel (S&P 500 + Nifty + curated watchlists); symbols need >=300 bars.
How was the Candlestick Pattern Scanner tested?
Tests surviving BH correction (2026-07-15 audit): 0 of 15; Edge range vs matched control: -1.9pp to +1.3pp; Hammer / Three White Soldiers: measured INVERTED vs claimed direction.