The OOS verdict flagged a variant 'candidate to promote' on hit-rate >= 0.5+margin
+ lag_stable alone. On thin data this over-claims: Source A Almon-ADL scored 6/10
(0.60) lag-stable and was flagged as signal, but P(X>=6|10,0.5)~=0.377 -- a coin
flip. Live ground-truth confirmed no signal (full-sample R2~=0.003, wrong sign).
Add exact stdlib-only one-sided binomial _binom_sf_ge + _VERDICT_ALPHA=0.05 and
require P(X>=hits|n_test,0.5) < alpha in both verdict() and cross_source_verdict()
on top of the effect-size margin. hits recovered exactly as round(hit_rate*n_test)
(n_test==scored invariant; no evaluator shape change). Verdict text now states
n_test + the binomial p on pass and fail. Evaluator/estimator math and the
read-only SELECT discipline untouched. Refs #978.
Extend the read-only §9.6 rate-sensitivity OOS harness with two opt-in
candidate-method variants so any wiring decision is evidence-based:
- --almon: evaluate_oos_almon, Almon distributed-lag (regression.fit_almon_dl),
fit on TRAIN only, point-in-time sum_j beta_j*drate[t-j] prediction.
- --deseasonalize: train-only month-of-year factors (normalize.seasonal_factors)
divided out before log_diff, then the existing best_lag evaluator.
Both pin the fit to _time_ordered_split(n_train); no look-ahead leakage
(adversarial tests assert the train fit is byte-identical under test corruption).
Default path (best_lag/raw) is byte-identical to before. 88 tests pass, ruff clean.
Prod OOS findings (directional hit-rate, coin-flip 0.50, bar 0.55+lag-stable):
- #979 deseasonalize: neutral (B 0.148->0.148, A 0.40->0.40) -> keep advisory.
- #978 Almon-ADL: dominates best_lag (B 0.148->0.407 lag-stable; A 0.40->0.60,
clears coin-flip+margin) -> candidate to promote from advisory.
REOPENED 951-B §9.6.
PART A: fix look-ahead leakage in backtest_rate_sensitivity --detrend. The
ln(units) trend was fit over train+test then split, so test data shaped the
detrend and inflated the OOS hit-rate. _detrend_log now takes fit_n; backtest_tier
fits the trend on TRAIN months only (same split evaluate_oos uses) and projects
(a,b) point-in-time onto test. Default fit_n=None preserves prior behaviour.
PART B (DoD): new app/services/forecasting/regression.py — Almon polynomial
distributed-lag (deg 2) of Δln(district demand) on Δkey_rate lags 0..6 via
OLS-on-Almon-regressors (numpy lstsq) + per-lag reconstruction + manual
Newey-West HAC SEs (NO statsmodels). Output {best_lag_months, coef=long-run
multiplier, x_pct, r2, n, per_lag_coef, hac_se,...}; gate mirrors _elasticity_coef
(n<30 OR R²<0.1 OR Σβ≥0 → fallback); §9.6 phrase from the lag shape. ADVISORY,
shipped standalone (integration point documented), NOT wired — protects the live
compute_rate_sensitivity consumers.
125+31 tests (synthetic known-lag recovery, HAC computed/differs-from-OLS,
fallback gating, no-leakage detrend). ruff clean. Refs #978
