Berg volume/TRIN/capitulation indicator family — capitulation proxy
Rejected qm-tykd.2 · 2026-08-17
Do Berg volume/TRIN/capitulation signals (volume climaxes, TRIN/Arms Index extremes, down-volume dominance, selling-pressure extremes) predict turning points beyond unconditional and event-appropriate baselines under causal lag/costs, and does prediction become a tradable overlay?
No. Primary SPY 1-day return <= -2.5% capitulation proxy (183 signals 1993-2026, 2.17% sparse) beats unconditional overall (mean fwd20 T+1 2.03% vs 0.92%, delta +1.12pp, hit 67.2% false 32.8% vs event -1.5% to -2.5% 1.15% delta +0.88pp; event -1% to -2.5% same) and retains at T+2 (+0.17pp to 2.20%), but fails frozen robustness: 2008-2019 delta -0.24pp hit 59.7% false 40.3% (regime failure vs 1994-2007 +3.19pp hit 75.8% and 2020+ +0.43pp), adjacent thresholds fragile (-3.0% drops to +0.51pp vs primary +1.12pp; no stable plateau; -1.5% +0.51pp, -2.0% +0.83pp, -4.0% +1.11pp volatile n=48), data alternatives too frequent or mismatch (VIX>=25 17.9% freq +1.04pp, VIX>=30 8.3% +1.55pp; down 4/5 days 12.3% +0.75pp; down 5/5 sparse n=93 +1.12pp), overlay 20d hold CAGR 3.55% vs SPY 10.74% delta -7.19pp (22.9% exposure 120 transitions) and share-volume TRIN/down-volume/volume_z lanes insufficient_history offline (Volume=0 close_only).
Provenance
Equity vs S&P (aligned, rebased to 1.0)
Aligned window, both rebased to 1.0 at start — Aligned window, both rebased to 1.0 at start for fair comparison · S&P dashed, experiment solid. X = Date, Y = Equity (rebased to 1.0).
Signal driver & thresholds (x-aligned)
Signal timeline (sparse triggers)
Triggers when overlay takes over base strategy
Stats
| CAGR | 0.00% | Sharpe | 0.00% |
|---|---|---|---|
| Sortino | 0.00% | maxDD | 0.00% |
Disposition
Rejected — capitulation proxy shows regime-dependent hit-rate that does not survive subperiod stability, adjacent-threshold robustness, or portfolio translation; share-volume TRIN/down-volume/volume climax require licensed NYSE volume (insufficient_history offline); no robust indicator to promote, no tradable strategy queued. All false capitulations and negative results retained.
Follow-on work
- No explicit follow-on; file new issue if needed
Errors & data gaps
NYSE TRIN (adv_vol/dec_vol)/(adv_issues/dec_issues), NYSE down-volume % (declining_volume/total_volume), and NYSE/SPY volume climax z-score/pct requiring share volume are insufficient_history offline - data/market.sqlite is close_only (Volume=0 for all ohlcv rows per qm-zl9.1/qm-4bx.11; 0 rows with volume>0) and no NYSE volume/TRIN table exists; proxy mismatch retained; true NYSE volume/TRIN test requires licensed exchange history via qm-zfl5/qm-m1ib and is documented as untestable missing input Exact Berg volume/TRIN calculations, TRIN thresholds (e.g., 1.5/2.0/2.5), down-volume dominance ratios (75/80/90%), lookbacks (1/5/10d, 20/63d) and publication lags withheld across all three reports (zero verbatim numeric thresholds) — all thresholds/windows reported are inferred price-derived proxy neighborhoods (ret1 <= -2.5% primary etc.) kept distinct from verbatim disclosure Offline proxy is price-derived selling-pressure (SPY return <= -2.5%) plus count-derived down-day and VIX spike proxies, not true volume - volume proxy correlation not assumed; 30k-indicator multiple-testing hazard severe even for modest overall +1.12pp edge with ~30 implicit threshold/window/alternative trials Sparse signal (2.17% frequency) — small subperiod ns (62 in 1994-2007, 77 in 2008-2019, 43 in 2020+) widen confidence; regime non-stationarity dominates (2008-2019 negative delta) Equity curve is illustrative 20d holding overlay (exp=1 within 20d post-capitulation, T+1 lag, fee_bps=10) vs SPY buy-and-hold, not a recommended allocation — prediction vs portfolio kept separate per acceptance; overlay underperforms by 7.19pp (3.55% vs 10.74% CAGR) with 22.9% exposure
Follow-up
gh issue create --title "Follow-up Berg volumeTRINcapitulation indicator family capitulation p" --body "Follow-up to qm-tykd.2: Do Berg volume/TRIN/capitulation signals (volume climaxes, TRIN/Arms Index extre" --label "priority:p2" --label "pipeline:research" --label "agent:ready"