← Back to dashboard

Cool-off / Overextended Cohort Tracking

Generated 2026-05-19T22:22:52Z · as_of 2026-05-19 · Overall: OK

Summary: clean_go=31 · overext_bb=2 · weak_go=10 · snaps=10 · 5d clean=+0.30%/overext=-0.32% (n=110/50) · rec=keep_advisory
Caveat: DIAGNOSTIC ONLY — production scoring formulas are NOT mutated. Tracks per-ticker forward returns of Pine cohorts to test whether overextended_bb names underperform clean-go names before promoting the blocker into LOW_RISK / SWING gating.
Recommendation: keep_advisory
5d gap is +0.62pp (clean=+0.30%, overext=-0.32%, n=110/50); not large enough to justify scoring change.

Current cohort sizes

CohortSizeMembers
clean_go31AFL, AIZ, ARMK, AVT, BK, BWA, CACC, CPAY, DTM, DUOL, EOG, EXEL, FRT, FSLR, GEN, GOOG, GWW, JEF, MDB, MTSI, MU, NDAQ, NVDA, OHI, OKE (+6 more)
overextended_bb2COST, D
weak_go10AME, CIEN, CNM, CVNA, ECG, EME, EQIX, ICE, RS, SCCO

Forward-return comparison by horizon

Each observation is one (snapshot, ticker) pair where the horizon has elapsed (no lookahead). Pending horizons are the count of snapshots whose elapsed trading days are still below the horizon.

HorizonCohortn obsmeanmedian%+snaps done / pending
1dclean_go326-0.21%0.00%42.02%9 / 1
1doverextended_bb69-0.39%-0.39%40.58%9 / 1
1dweak_go49-0.45%-0.25%28.57%9 / 1
3dclean_go195-0.36%-0.25%46.67%6 / 4
3doverextended_bb581.26%1.10%60.34%6 / 4
3dweak_go26-1.46%-1.26%30.77%6 / 4
5dclean_go1100.30%0.44%56.36%4 / 6
5doverextended_bb50-0.32%0.26%56.00%4 / 6
5dweak_go17-3.90%-2.80%23.53%4 / 6
10dno completed snapshots yet0 / 10

Snapshot scaffold

Snapshots tracked: 10 (retention 90 days, cap 500 rows). Min observations before promoting blocker into scoring change: 30 per cohort at 5d.