SMH/SPY · > 5% below its 65-day MA · forward 21 sessions · recency window from 2021-01-01 · half-life 9mo · de-clustered ≥42d
Now: -8.1% vs the 65-day MA · 4th pctile (2021+) · 2th all-time
Close 0.72 · MA 0.79 · IN BREACH now · all-time range -12.7% … +25.6%
Forward returns by regime
Median forward return (% positive) at 2 / 5 / 8 / 11 / 15 / 21 sessions, conditioned on each market regime — and the recency-weighted read (the headline). Watch how the shape shifts as the window tightens toward now.
Full historyPost-CovidAI eraRallyRecency-wtdbaseline
Horizon
Full historyn=32
Post-Covidn=11
AI eran=4
Rallyn=4
Recency-wtdn=11
2 sessions
-0.1% 47%
-0.5% 36%
-1.2% 25%
-1.2% 25%
-6.9% 12%
5 sessions
+0.7% 58%
+1.4% 60%
-1.6% 33%
-1.6% 33%
-1.6% 41%
8 sessions
+1.3% 61%
+0.4% 50%
-0.4% 33%
-0.4% 33%
-0.4% 38%
11 sessions
+1.2% 65%
+1.9% 80%
+2.7% 100%
+2.7% 100%
+2.7% 94%
15 sessions
+2.5% 65%
+2.9% 80%
+2.7% 100%
+2.7% 100%
+2.7% 94%
21 sessions
+1.9% 68%
+3.3% 70%
+1.9% 67%
+1.9% 67%
+1.9% 59%
Regime delta: the recency-weighted 21-day median is +0.1 pts vs the full-history read — little regime change.
Headline is recency-weighted (faster 9-month half-life) over 11 post-2021 events. Regime windows: AI era from Apr 1 2023 · Rally from the Oct 30 2023 bottom. Tight regimes carry few events (see the n per column) — read them with the recency-weighted line, which keeps sample by fading rather than cutting.
Forward 21d · median
+1.9%
mean +1.6%
% positive
59%
11 events
vs baseline
+1.2 pts
base +0.7% / 56% pos
Range
-3% … +13%
25th -0.1% · 75th +8.1%
Path · max-gain
+5.1%
median MFE in window
Path · max-draw
-1.8%
median MAE in window
Every event · outcome over time
Each dot is one de-clustered event at its date; height is the forward 21-session return. Gold-ringed = post-2021 — if there are few gold rings, the signal is a pre-Covid-regime artifact.
By decade: 2000s 112010s 102020s 10
The event log
Event date
Stretch
Fwd 21d
Max gain
Max draw
2005-10-28
-6.5%
+5.8%
+6.1%
-0.6%
2006-03-09
-5.4%
+0.6%
+1.7%
-3.9%
2006-05-25
-5.7%
-2.1%
+1.0%
-3.4%
2006-07-26
-9.7%
+5.3%
+6.0%
+0.0%
2007-10-11
-6.0%
-5.7%
+0.0%
-8.6%
2008-01-02
-5.5%
+0.1%
+0.1%
-7.0%
2008-07-30
-5.8%
+3.2%
+8.9%
+0.0%
2008-09-29
-5.7%
+2.4%
+8.9%
-4.9%
2008-11-26
-8.5%
+0.9%
+7.4%
-3.6%
2009-05-08
-7.7%
+9.6%
+10.2%
+0.0%
2009-11-03
-5.8%
+6.2%
+6.2%
+0.0%
2010-08-13
-5.4%
-1.7%
+1.9%
-5.1%
2011-07-13
-5.5%
+1.0%
+1.7%
-1.8%
2012-05-24
-5.7%
-0.2%
+2.5%
-0.3%
2012-09-24
-5.9%
-0.3%
+0.9%
-2.7%
2014-10-10
-6.7%
+6.5%
+7.5%
+0.0%
2015-07-09
-6.2%
-2.3%
+0.7%
-3.5%
2018-04-23
-5.8%
+6.9%
+6.9%
+0.0%
2018-08-17
-6.2%
+1.0%
+5.0%
-0.3%
2018-10-18
-5.5%
-1.3%
+2.6%
-4.1%
2019-05-20
-6.9%
+1.9%
+4.0%
-1.4%
2021-03-08
-7.0%
+9.5%
+11.4%
+0.0%
2021-05-06
-5.9%
+3.4%
+3.9%
-3.8%
2022-01-27
-7.2%
+3.2%
+5.8%
-0.5%
2022-04-05
-6.9%
-0.7%
+0.4%
-4.9%
2022-07-01
-9.7%
+13.0%
+13.2%
+0.0%
2022-09-01
-5.8%
-1.6%
+0.7%
-2.6%
2022-11-01
-6.7%
+12.5%
+14.3%
-0.5%
2024-07-30
-8.3%
+1.9%
+6.1%
-3.7%
2024-11-26
-5.4%
+4.0%
+4.4%
-1.0%
2025-02-27
-5.4%
-2.7%
+3.2%
-2.7%
Read it honestly. The recency-weighted read leans on 11 post-2021 events (half-life 9mo); the full-history column carries 32. The regime delta shows how far recent behaviour has drifted from the long base rate. De-clustered by a 42-session gap; overlapping windows remain. A context gauge, not a mechanical signal.
Computed live from adjusted-close history (yfinance). Forward return = close-to-close over 21 sessions from each event. MFE/MAE = max favorable / adverse excursion within the window. Educational research, not investment advice.