Plan: #180 merged; the quantile queue holds (I076)

benchmarks/CAMPAIGN_PLAN.md

@@ -186,7 +186,7 @@ fact: 2026-09-25 | standing quantile BASE = `results/quantile-20260925-172250.js
 | F3 | Classifier forced-cross | KILLED 2026-09-21 (S3, I021) | gr binary engaged 10W-13L, median −0.04%: the race earns its fee on the classifier. Knob stays opt-in (PR #117), no rung-1 pin |
 | F5 | hc-Brier gap vs CatBoost | **SHIPPED 2026-09-22 (I046, PR #145, d14bf38): `cat_count_features` on by default** (public 5W-1L, decide hc 6W-1L, bit-identical elsewhere, chart refreshed) — earlier: I038 library form, I039 S3 PASS | the published chart (`public_pareto.png`) refresh: DONE 2026-09-24 on 0.33.0 (`results/20260924-090541.json`, `--public --seeds 3`, 22 datasets, ~4 h: the default avg rank 1.94 [1.64-2.24] against CatBoost 1.84 [1.43-2.24] and LightGBM 2.23, at a median 5.6x against 47.2x and 1.0x; `docs/benchmarks.md` prose updated; `pareto.png` re-rendered from `results/20260924-130232.json` and `docs/PROJECT_STATUS.md` synced); S7 (multiclass CTR width) is the family's next idea if picked. History: **A per-categorical count column closes 47% of CatBoost's hc edge** (I035): 4W-1L on the gap sets, +0.43% Brier median, gains ordered by cardinality, sf-police and Traffic unanimous across seeds. The gap is the encoder (CatBoost on our TS keeps none of its edge); not the prior target, Counter, permutations or quantization (TS quantization kills on big sets, +1.7–3.6% on the two small controls — a small-data pointer, parked). `cat_count_features` (opt-in, card ≥ 256, invisible to the cross and linear-leaf races; I038) on the decision tier (I039): gr 0-0-59 exact ties, the 7 hc sets without a qualifying column exact ties, the engaged 7 **6W-1L** at +0.20% median (sf-police +0.73%, Traffic +0.86% Brier; employee_salaries +2.45%, wine-reviews +0.56% RMSE), hc@time 4-0, fit ×1.09 on hc (engaged median 1.165). PR up with the flag OFF. The random-effects alternative (per-column ANOVA λ for the TS, I040) KILLED: uncapped it collapses the small controls (−3.8 / −9.6%), capped at 10 it is a flat wash and still costs kick and eucalyptus; the count column keeps evidence the shrinkage deletes. Next: the maintainer's go on /experiment S4 for the default flip; meanwhile R4 S0 |
 | F6 | Ordered-TS train/test moment mismatch (shortlist R1) | KILLED 2026-09-21 (S1b, I025) — closed as barrier B18 | The defect is real (rare categories over-trusted, reliability 0.63–0.70) and two transform-side fixes both went 7W-5L against a bar of 8: the gain is sf-police (9 of 9 fits, +0.29% to +0.53%) and nothing else. Nothing ships; the open door is the Counter feature, which belongs to R3 |
-| F7 | Multi-quantile head (`ChimeraBoostQuantileRegressor`) | ACTIVE again 2026-09-25 (the maintainer: "we can continue on the quantile thread"; PARKED 2026-09-24 for GitHub issues; ACTIVE from 2026-09-23). Phase 1, the bench, MERGED (I056, PR #156); Q0 DONE (I057, PR #157 merged): uncapped and validation-rescaled pay, depth 6 misses on its guard alone, recentred fails. Q3 CLOSED (I058, barrier B23: a larger rate loses); depth 6 + early-stopping-row calibration PAID (31W-5L, +0.33%, 90% error 3.43 → 0.47); PR #158 merged. **Q5 PASSED (I059): the head's default is now depth 6 + `conformalize="auto"`** (gr 31W-5L, +0.33%, 90% coverage error 3.43 → 0.47); PR #159 merged, snapshot rebaselined. **Q2 CLOSED (I060)**: no probe clears the tightened gate (sign p < 0.05); the five low-noise sets need resolution (bins) on two and a squared-error location on two; PR #160 merged. **Q6 T0 PASSED (I061)**: the validation-chosen head wins 23W-7L-6T (p 0.005), recovers 99% of the oracle, and tops the 59-key frontier above CatBoost MQ (0.6006 @ 7.2× against 0.5982 @ 129×); PR #161 merged. **Q7 PASSED (I062): the audition is the head's default** (identity 177/177 with the bench arm; gr 23W-7L-6T against the Q5 default, p 0.005; CatBoost MQ now 9W-27L against us, off the frontier); PR #162 merged, **released in 0.33.0**. **Q8 T0 PASSED (I070)**: a fixed-width (S) and a scaled-residual (N) audition candidate, gr 18W-1L-17T against the Q7 default (p 7.6e-5) at 1.11x the fit; S alone and uncapped rounds fail. **Q9 PASSED (I071): S and N in the library default** (identity 177/177; CatBoost MQ 29W-7L, RigidShift 35W-1L; the 59-key chart 0.6046 @ 8.3x); merged as PR #177 (a6a90d2); flag: hc `@time` 90% coverage error 2.12 -> 4.69 (Moneyball, a pointer). **Q10 T0 (I072) DIAGNOSED**: RoNGBa's lead on visualizing_soil and SGEMM is the S/N centre model's accuracy (our grid on its mean beats it); pol is a width gap. **Q11 T0-a (I073)**: centre settings on five sets: 254 bins matches RoNGBa's lead (visualizing_soil +24%, SGEMM +7%) but costs cpu_act 2.5%; 8000 rounds +5% on visualizing_soil only; bagging +30% (ceiling, not a default). **T0-b (I073b) FAILS**: as extra candidates the 254-bin centre still costs cpu_act 2.5% (validation cannot see the overfit); 8000 rounds closed unrun (the centre is cap-bound on 1 of 59 keys). The RoNGBa line is CLOSED. **Q4 CLOSED (I074)**: the N candidate already is the spread-aware categorical encoding; `catscale` at 10k now 16.0 against CatBoost MQ's 30.9 (excess CRPS x1000). **I063 DONE (PR #167)**: RoNGBa is the NGBoost opponent (issue #163) | next: **Q13 (I076) PASSED, PR for the maintainer**: `refit_full=True` as the head's default (gr 35W-1L, +0.91%, p 1.1e-9, 1.28x fit); after the merge, Q1 (P16). Program: `QUANTILE_PLAN.md` "Campaign 2026-09-23" |
+| F7 | Multi-quantile head (`ChimeraBoostQuantileRegressor`) | ACTIVE again 2026-09-25 (the maintainer: "we can continue on the quantile thread"; PARKED 2026-09-24 for GitHub issues; ACTIVE from 2026-09-23). Phase 1, the bench, MERGED (I056, PR #156); Q0 DONE (I057, PR #157 merged): uncapped and validation-rescaled pay, depth 6 misses on its guard alone, recentred fails. Q3 CLOSED (I058, barrier B23: a larger rate loses); depth 6 + early-stopping-row calibration PAID (31W-5L, +0.33%, 90% error 3.43 → 0.47); PR #158 merged. **Q5 PASSED (I059): the head's default is now depth 6 + `conformalize="auto"`** (gr 31W-5L, +0.33%, 90% coverage error 3.43 → 0.47); PR #159 merged, snapshot rebaselined. **Q2 CLOSED (I060)**: no probe clears the tightened gate (sign p < 0.05); the five low-noise sets need resolution (bins) on two and a squared-error location on two; PR #160 merged. **Q6 T0 PASSED (I061)**: the validation-chosen head wins 23W-7L-6T (p 0.005), recovers 99% of the oracle, and tops the 59-key frontier above CatBoost MQ (0.6006 @ 7.2× against 0.5982 @ 129×); PR #161 merged. **Q7 PASSED (I062): the audition is the head's default** (identity 177/177 with the bench arm; gr 23W-7L-6T against the Q5 default, p 0.005; CatBoost MQ now 9W-27L against us, off the frontier); PR #162 merged, **released in 0.33.0**. **Q8 T0 PASSED (I070)**: a fixed-width (S) and a scaled-residual (N) audition candidate, gr 18W-1L-17T against the Q7 default (p 7.6e-5) at 1.11x the fit; S alone and uncapped rounds fail. **Q9 PASSED (I071): S and N in the library default** (identity 177/177; CatBoost MQ 29W-7L, RigidShift 35W-1L; the 59-key chart 0.6046 @ 8.3x); merged as PR #177 (a6a90d2); flag: hc `@time` 90% coverage error 2.12 -> 4.69 (Moneyball, a pointer). **Q10 T0 (I072) DIAGNOSED**: RoNGBa's lead on visualizing_soil and SGEMM is the S/N centre model's accuracy (our grid on its mean beats it); pol is a width gap. **Q11 T0-a (I073)**: centre settings on five sets: 254 bins matches RoNGBa's lead (visualizing_soil +24%, SGEMM +7%) but costs cpu_act 2.5%; 8000 rounds +5% on visualizing_soil only; bagging +30% (ceiling, not a default). **T0-b (I073b) FAILS**: as extra candidates the 254-bin centre still costs cpu_act 2.5% (validation cannot see the overfit); 8000 rounds closed unrun (the centre is cap-bound on 1 of 59 keys). The RoNGBa line is CLOSED. **Q4 CLOSED (I074)**: the N candidate already is the spread-aware categorical encoding; `catscale` at 10k now 16.0 against CatBoost MQ's 30.9 (excess CRPS x1000). **I063 DONE (PR #167)**: RoNGBa is the NGBoost opponent (issue #163) | next: **Q13 (I076) MERGED (PR #180)**: `refit_full=True` is the head's default (gr 35W-1L, +0.91%, p 1.1e-9, 1.28x fit). F7 HOLDS for the maintainer's call: Q1 (P16) re-read as low value (I076). Program: `QUANTILE_PLAN.md` "Campaign 2026-09-23" |
 
 ### F1 — Cross-feature cost trim v2
 status: KILLED 2026-08-16 at S2 (I007 at k=6, I008 at k=12) — closed as barrier B16
@@ -495,6 +495,17 @@ forecast: (1) HIT; (2) HIT; (3) HIT; (4) HIT (I075's numbers, reproduced
 exactly).
 verdict: **PASS -> PR for the maintainer** (library, tests, docs). After
 the merge: rebaseline the identity snapshot, delete the branch.
+MERGED 2026-09-25 as PR #180 (180f765); branch deleted; the identity
+snapshot re-checked (182/186, only `mq3` and `mq3_w_sub`) and rebaselined
+at 180f765 (186/186).
+next (F7), re-read 2026-09-25: Q1 (P16, honest leaf quantiles) is the only
+queued item, and its own draft calls the bias-variance trade "genuinely
+uncertain"; it needs a third split or a per-round partition and breaks a
+recorded booster invariant, and its coverage motivation is gone (Q5). The
+"leaf refit cost" item is not a cheap win either: each of the K channels
+selects over its own residual vector, and three optimizations already
+landed (this file's speed note). F7 holds for the maintainer's call on
+what comes next.
 
 #### I075 2026-09-25 F7 Q12 T0 (the head retrains its winner on all rows after the audition, as the point regressor's `refit_full` does; BENCH-only probe arms, pre-registered)
 why now: the maintainer, 2026-09-25, picking option 1 of three: "1 is

benchmarks/QUANTILE_PLAN.md

@@ -436,7 +436,7 @@ Q0 reads moved it down to item 4.
    20W-1L-15T (p 2.1e-5), so (b). Coverage guard fine (gr 90% error 0.32 →
    0.47 points); the hc `@time` flag shrinks (4.69 → 3.29).
 3h. **Q13, `refit_full=True` as the default.**
-   **PASSED 2026-09-25 (I076), PR for the maintainer.** Without an
+   **PASSED 2026-09-25 (I076); merged as PR #180 (180f765).** Without an
    `eval_set` the default reproduces Q12's arm (b) bit for bit; with one,
    nothing changes. The suite's field arm keeps its `eval_set`, so every
    comparison stays "every model on the same rows" and leaves this gain