Verdict: Intraday Trade Ranker (M-030) is rated “NOT FIT (D1h/D1/D0 direction rankers)” in its validation review dated 4 Nov 2026.
Key result: Hold-out IC 10:30 ET (2026, 209 sessions): +0.015, NW t 1.8.
Key result: Hold-out LS-10 net 10:30 ET: −22.8 bp/day, t −3.9; gross −3.8.
How it was tested: training data 2023-01→2025-12, test data 2026-01→2026-10; out-of-sample design (how it was checked on data it never saw while being built): not stated.
Horizon: Same day (10:30 ET → close). Coverage: US stocks ≥ $10M/day and ≥ $5 (~2,300 names), 2023 → (intraday survivorship).
Forward record: 2 of 3 published lists have reached their 1-session horizon. Average pick return minus SPY's over the same dates: +0.03 percentage points (standard error 1.25). 48% of the matured picks beat SPY over the horizon.
Stage: Validated — NOT FIT (shadow only). Version v1; the hub record is as of 6 Nov 2026.
The verdict
NOT FIT for trading; replication exact, no leakage found; the range model is fit for stop/size guidance only.
The validation record also states the rating as “NOT FIT (D1h/D1/D0 direction rankers) · FIT WITH CONDITIONS (V1 range model, stops/size only)”.
Development summary: NOT FIT: no intraday direction edge survives costs at 09:30, 10:00 or 10:30 ET in pre-registered hold-outs; the range model is a modestly sharper ATR for stops and size.
What it predicts
Item
Detail
Question it answers
Which liquid US stocks to day-trade after the first hour: ranked at 10:30 ET, long/short, exit at the close. Shadow only.
Method
Ridge regression
How it works
Ridge regression on 72 within-day ranks (champion of a fixed 15-model grid: ridge, boosted trees, MLP)
Inputs
Yesterday's return (reversal), Opening-range return vs sector, Opening-range candle (close location), Distance from day's high/low (LoD%), Opening-range relative volume
Output
Daily top/bottom-10 shadow list at 10:30 ET with stop (LoD/HoD), expected range, settled at the close
Horizon
Same day (10:30 ET → close)
Coverage
US stocks ≥ $10M/day and ≥ $5 (~2,300 names), 2023 → (intraday survivorship)
Markets
US
Model type
Model · ML/Stat
Stage
Validated — NOT FIT (shadow only)
Version
v1, first recorded 3 Nov 2026
Validation evidence
Training data: 2023-01→2025-12. Test data: 2026-01→2026-10. Out-of-sample design: not stated.
Headline results from the model's development record, as the hub publishes them.
Measure
Result
Compared with
Note
10:30 ET LS-10 net, hold-out 2026
−22.8 bp/day
gross −3.8 · R1 −14.9 · R2 −16.2 · R3 −13.6
Daily IC, hold-out 2026
+0.015 (t 1.8)
dev 2023–25 +0.019 (t 2.9)
Open-print edge by entry time
+24.7 → +6.8 bp
open → 09:35 on the same 64 sessions; 10:30 −27.7
Published 5-min breakout, 72 sessions
−28.9 to −2.0 bp
conservative vs optimistic fills
Range model R², hold-out
0.576
ATR% + opening range 0.555; 9/10 months better
Forward record: 2 of 3 published lists have reached their 1-session horizon. Average pick return minus SPY's over the same dates: +0.03 percentage points (standard error 1.25). 48% of the matured picks beat SPY over the horizon.
Limitations
Intraday history 2023 → only, survivors only
Hourly/30-min bars; 5-min only for ~72 sessions
Questions
Is the Intraday Trade Ranker validated?
Its validation review dated 4 Nov 2026 rated it NOT FIT for trading; replication exact, no leakage found; the range model is fit for stop/size guidance only.
What does the Intraday Trade Ranker predict?
It answers: Which liquid US stocks to day-trade after the first hour: ranked at 10:30 ET, long/short, exit at the close. Shadow only. Output: Daily top/bottom-10 shadow list at 10:30 ET with stop (LoD/HoD), expected range, settled at the close. Horizon: Same day (10:30 ET → close). Coverage: US stocks ≥ $10M/day and ≥ $5 (~2,300 names), 2023 → (intraday survivorship).
How was the Intraday Trade Ranker tested?
Training data 2023-01→2025-12, test data 2026-01→2026-10, out-of-sample design: not stated. Hold-out IC 10:30 ET (2026, 209 sessions): +0.015, NW t 1.8; Hold-out LS-10 net 10:30 ET: −22.8 bp/day, t −3.9; gross −3.8; Open-print model LS-10 net (2025-07→2026-10): +26.7 bp/day at the print; +6.8 at 09:35; −27.7 at 10:30. Forward record: 2 of 3 published lists have reached their 1-session horizon. Average pick return minus SPY's over the same dates: +0.03 percentage points (standard error 1.25). 48% of the matured picks beat SPY over the horizon.