Verdict: US 12-month Sector-Routed Ranker (M-027) is rated “NOT FIT” in its validation review dated 19 Oct 2026.
Key result: 12m within-sector rank IC 2021-24: +0.119 (momentum +0.029).
Key result: Gap over 12-1 momentum: +0.090, t +1.32, not significant.
How it was tested: training data 2012-01→2019-12, test data 2021-01→2024-12; out-of-sample design (how it was checked on data it never saw while being built): not stated.
Stage: Validated — NOT FIT (shadow descriptive output). Version v1; the hub record is as of 6 Nov 2026.
The verdict
NOT FIT as a 12-month ranker: it does not beat 12-1 momentum, its own IC fails placebo calibration, and its sign flipped across 2012-19, 2021-24 and 2025.
Development summary: Registered rule -> NOT FIT. V_EQ 12m within-sector IC +0.119 (NW t +2.94) beats E/P and M-001 but not 12-1 momentum (+0.090, t +1.32); fitted challengers lose to it.
What the review allows
rank_pct as a DESCRIPTIVE within-sector read of cheapness + quality + low volatility (what the composite says, not a forecast)
components to see which input drives a name's position
What the review rules out
Treat rank_pct as an expected-return or 12-month outperformance forecast
Claim a per-sector edge: placebo cells pass the frozen inference
Prefer it to momentum or trailing P/E as a stock selector
What it predicts
Item
Detail
Question it answers
Rank US stocks within their sector for a 12-month hold, routing the value read by business model (banks, bond-proxies, everything else)
Method
Routed rank composite (V_EQ)
How it works
Equal-weight within-sector rank composite: routed value + quality + low volatility; per-route OLS and gradient-boosting challengers
Training data: 2012-01→2019-12. Test data: 2021-01→2024-12. Out-of-sample design: not stated.
Headline results from the model's development record, as the hub publishes them.
Measure
Result
Compared with
Note
12m pooled IC, test 2021-24
+0.119
MOM +0.029
NW t +2.94; calibrated t +1.49
V_EQ minus 12-1 momentum
+0.090
0
t +1.32, 2/4 years > 0
Top-bottom decile, 12m
+2.3 pp
0
t +0.28
Train-window IC 2012-19
+0.018
test
decile spread -12.2 pp
2025 forward read
-0.140
test
8 formations, not measurable
The hub's standard metrics for this model class (only those the record measured).
Metric
Sample / window
Compared with
Value
Range (CI)
Rank IC (headline)
2021-01→2024-12
0.0287
0.119
n/a
ICIR
2021-01→2024-12
n/a
1.11
n/a
Decile spread
2021-01→2024-12
0
+2.34 pp
n/a
Limitations
Persistent scores inflate Newey-West t about 2x (placebo kappa 1.98); sector cells are not supportable.
Base panel misses some US filers and ~4 in 5 delistings.
Questions
Is the US 12-month Sector-Routed Ranker validated?
Its validation review dated 19 Oct 2026 rated it NOT FIT as a 12-month ranker: it does not beat 12-1 momentum, its own IC fails placebo calibration, and its sign flipped across 2012-19, 2021-24 and 2025.
What does the US 12-month Sector-Routed Ranker predict?
It answers: Rank US stocks within their sector for a 12-month hold, routing the value read by business model (banks, bond-proxies, everything else). Output: Per US name: within-sector percentile (100 = most attractive), score, top-5 components. Horizon: 12 months (252 sessions) from a month-end formation; 6 months secondary. Coverage: deep-value-lab PIT panel eligible names, mcap >= $300M, >= $1M/day, >= $1, sector known; 93 delisted names carried.
How was the US 12-month Sector-Routed Ranker tested?
Training data 2012-01→2019-12, test data 2021-01→2024-12, out-of-sample design: not stated. 12m within-sector rank IC 2021-24: +0.119 (momentum +0.029); Gap over 12-1 momentum: +0.090, t +1.32, not significant; Placebo-calibrated t (12m): +1.49 — not significant.