red since 2025-10-17 · now 0.83 pp · macro layer · weight 2.0
Rule: amber: inverted (<0). red: re-steepened above 0 after an inversion in the last 12 months
Track record: Inverted before every US recession since 1969; equity peaks came 0-18 months after inversion (2000, 2007). The re-steepening is usually the closer warning.
amber since 2026-10-16 · now 19.99 VIX · market layer · weight 1.5
Rule: amber: VIX >= 20 or term ratio >= 0.95. red: VIX >= 30 or ratio >= 1 (backwardation)
Track record: Backwardation (spot above 3-month) appeared in Aug 2015, Feb 2018, late Feb 2020 and early Apr 2025; it is the clearest same-week signal of a forced-selling crash.
Moderate risk · 28/100 Tighten stops, trim the weakest/most speculative positions, start a buy list.
Composite
28
Macro layer
31
Market layer
21
S&P from 52w high
-2.0%
What to do now
Posture (Moderate): Tighten stops, trim the weakest/most speculative positions, start a buy list.
Hedge cost: VIX 20.0: puts are normally priced. Outright puts or put spreads both work; spreads cut the cost if VIX is rising.
Capitulation signal: not active (last: 2020-04-24).
Macro layer = slow indicators with months of lead (2000, 2008-type bear markets). Market layer = fast stress gauges that react within days (1987, 2020, 2025-type shocks). Capitulation buy setups (blue triangles), latest: 2002-07-29, 2008-12-17, 2009-04-03, 2020-04-24
Inverted before every US recession since 1969; equity peaks came 0-18 months after inversion (2000, 2007). The re-steepening is usually the closer warning.
macro
0.83 pp
red
since 2025-10-17
amber: inverted (<0). red: re-steepened above 0 after an inversion in the last 12 months
HY spreads bottomed in mid-2007 and widened for months while the S&P made its October 2007 high. Widening also flagged 1998, 2015-16, Q4 2018. Latest HY OAS: 3.23%.
macro
0.05 pp vs 6m low
green
since 2025-08-11
amber: +0.30pp above its 6-month low. red: +0.60pp
Claims bottomed in 2000 and 2006 ahead of the peaks; the Sahm rule confirms recession but usually after stocks have already topped (it is a confirmer, not an early signal).
macro
0.13 % above 52wk low
green
since 2026-02-02
amber: claims +12% off the low or Sahm >= 0.30. red: +25% or Sahm >= 0.50
Record highs in 2000 and 2021 set up -49% and -25% falls; but it stayed high for years (1997-2000, 2017-2021), so it scales risk rather than timing it.
macro
100.00 historical percentile
red
since 2020-04-01
amber: >= 75th percentile of its own history. red: >= 90th
Backwardation (spot above 3-month) appeared in Aug 2015, Feb 2018, late Feb 2020 and early Apr 2025; it is the clearest same-week signal of a forced-selling crash.
market
19.99 VIX
amber
since 2026-10-16
amber: VIX >= 20 or term ratio >= 0.95. red: VIX >= 30 or ratio >= 1 (backwardation)
Rolled over hard in 2008, 2011, 2015 and early 2020 as growth expectations collapsed.
market
-0.07 %
green
since 2026-04-10
amber: down 8%+ in 3 months. red: down 15%+
How tiles are sized
Risk score: 3 + composite / 20. Each indicator: its weight × (1 + 2 × state), so red tiles are 5× their green size and amber 3×. Safe havens: 1 + composite / 25. Put candidates: 0.8 + composite / 30. Led before past crashes: 0.8 + composite / 40. Next-crash regime picks: 1 + macro layer / 20. Rebound buy list: 0.8, jumping to 3 + 20 × (drawdown − 10%) once the S&P is 10%+ off its high, plus 6 while the capitulation signal is active. Backtest and crash study: 0.8 each. News: 1 + 0.4 × the number of high-relevance headlines in the last 24 hours (capped at 15), recomputed in the browser whenever news.json refreshes. Tiles are laid out as a squarified treemap with area proportional to weight.
Historical pattern ideas, not financial advice. Past crash behavior varies by cause.
Rule: an episode is inflationary / tightening if fed funds rose 1pp or more over the 12 months before the S&P peak, or the 10y yield rose at all over the 6 months before it; otherwise deflationary / recession. Today's readings pick the regime, and assets are ranked by their average peak-to-trough return in that regime's episodes.
Historical pattern ideas, not financial advice. Past crash behavior varies by cause.
red since 2020-04-01 · now 100.00 historical percentile · macro layer · weight 0.5
Rule: amber: >= 75th percentile of its own history. red: >= 90th
Track record: Record highs in 2000 and 2021 set up -49% and -25% falls; but it stayed high for years (1997-2000, 2017-2021), so it scales risk rather than timing it.
Every S&P decline of 15%+ since 1999 (6 episodes: 2000-03, 2007-10, 2018-09, 2020-02, 2022-01, 2025-02). Total returns from Yahoo adjusted closes; Treasuries, gold, oil and the dollar use older funds/futures before their ETFs existed. Assets need 4+ episodes to be ranked. Full table: sector_study.csv.
Asset
Avg
Median
Hit >0
Beat S&P
n
IEF Treasuries 7-10y
ETF IEF · futures ZN (10y T-note)
+9.3%
+4.9%
83%
100%
6
TLT Treasuries 20y+
ETF TLT · futures ZB (30y T-bond)
+9.2%
+9.4%
83%
83%
6
UUP US dollar
ETF UUP · futures DX (dollar index)
+6.0%
+3.8%
83%
100%
6
SHY Treasuries 1-3y
ETF SHY · futures ZF / ZT (5y / 2y notes)
+5.5%
+1.7%
83%
100%
6
GLD Gold
ETF GLD · futures GC (gold)
+3.9%
+1.6%
60%
100%
5
DBC Commodities
ETF DBC · futures broad commodity basket
-12.1%
-16.5%
20%
100%
5
Ranked by average return from S&P peak to trough. Index shorts or puts: ES / NQ futures, SPY / QQQ options.
Historical pattern ideas, not financial advice. Past crash behavior varies by cause.
green since 2025-08-11 · now 0.05 pp vs 6m low · macro layer · weight 2.0
Rule: amber: +0.30pp above its 6-month low. red: +0.60pp
Track record: HY spreads bottomed in mid-2007 and widened for months while the S&P made its October 2007 high. Widening also flagged 1998, 2015-16, Q4 2018. Latest HY OAS: 3.23%.
Today's 6-month sector leaders
XLK Tech +24.7% over 6m · ETF XLK · AAPL MSFT NVDA AVGO (large constituents, check current holdings)
XLV Health care +21.9% over 6m · ETF XLV · LLY UNH JNJ ABBV MRK (large constituents, check current holdings)
XLE Energy +19.2% over 6m · ETF XLE · XOM CVX COP (large constituents, check current holdings)
Check of the claim that pre-crash leaders fall hardest, sector ETFs only: correlation between the 6m-before return and the peak-to-trough return is -0.28 pooled and -0.11 average rank correlation within 6 episodes. The top 3 pre-crash sectors averaged -33.3% in the crash vs -25.3% for the bottom 3. The data weakly supports the claim.
Index alternatives: ES / NQ futures, SPY / QQQ puts. Puts are worth more when the risk band is Elevated or High.
Historical pattern ideas, not financial advice. Past crash behavior varies by cause.
Every S&P decline of 15%+ since 1999 (6 episodes: 2000-03, 2007-10, 2018-09, 2020-02, 2022-01, 2025-02). Total returns from Yahoo adjusted closes; Treasuries, gold, oil and the dollar use older funds/futures before their ETFs existed. Assets need 4+ episodes to be ranked. Full table: sector_study.csv.
Asset
Avg
Median
Hit >0
Beat S&P
n
QQQ Nasdaq 100
+29.9%
+16.7%
100%
100%
6
XLK Tech
+23.5%
+16.9%
100%
83%
6
XLY Discretionary
+12.3%
+13.5%
83%
50%
6
XLF Financials
+10.8%
+10.3%
100%
33%
6
XLV Health care
+10.2%
+12.6%
83%
17%
6
Ranked by average return in the 6 months before each S&P peak. These are candidates to hedge with puts when the risk band is Elevated or High.
Historical pattern ideas, not financial advice. Past crash behavior varies by cause.
green since 2026-02-02 · now 0.13 % above 52wk low · macro layer · weight 1.5
Rule: amber: claims +12% off the low or Sahm >= 0.30. red: +25% or Sahm >= 0.50
Track record: Claims bottomed in 2000 and 2006 ahead of the peaks; the Sahm rule confirms recession but usually after stocks have already topped (it is a confirmer, not an early signal).
green since 2026-04-08 · now 5.34 % from 200d · market layer · weight 1.5
Rule: amber: below the 200-day. red: below it and the 200-day is falling
Track record: Every bear market since 1929 spent most of its decline below a falling 200-day; staying out below it skipped most of 1973-74, 2000-02, 2008.
green since 2022-07-25 · now -0.50 index (0 = average) · macro layer · weight 1.0
Rule: amber: above -0.35 and up 0.15+ in 3 months. red: above 0 (tighter than average)
Track record: Crossed above zero in late 2007, 2020 and briefly 2022; a fast tightening is the channel through which Fed hikes and credit stress reach stocks.
green since 2024-02-05 · now -0.22 pp · macro layer · weight 1.0
Rule: amber: +1pp in 12 months. red: +2pp
Track record: Preceded 1987, 1990, 2000, 2007, 2018 and 2022. The Fed hikes until something breaks.
Loading news.json ...
Ranking: theme keyword weights (title hits count double), x1.5 + 1 for official sources (Fed, BLS, BEA, SEC), x1.25 / x1.5 when the linked indicator is amber / red, routine regulatory notices x0.3, and a 12-hour half-life. Score 3+ counts as high relevance. Feeds and themes: news.py. News refreshes hourly; this page re-checks every 15 minutes.
S&P -2.0% from high · capitulation not active. This list matters once the S&P is well off its high and the capitulation signal fires.
Every S&P decline of 15%+ since 1999 (6 episodes: 2000-03, 2007-10, 2018-09, 2020-02, 2022-01, 2025-02). Total returns from Yahoo adjusted closes; Treasuries, gold, oil and the dollar use older funds/futures before their ETFs existed. Assets need 4+ episodes to be ranked. Full table: sector_study.csv.
Trough +6 months
Asset
Avg
Median
Hit >0
Beat S&P
n
XLK Tech
+43.7%
+45.3%
100%
100%
6
XLF Financials
+40.0%
+23.0%
100%
33%
6
QQQ Nasdaq 100
+39.7%
+38.6%
100%
100%
6
XLI Industrials
+36.1%
+31.1%
100%
67%
6
XLB Materials
+35.1%
+22.6%
100%
67%
6
Trough +12 months
Asset
Avg
Median
Hit >0
Beat S&P
n
XLK Tech
+66.0%
+64.1%
100%
100%
6
QQQ Nasdaq 100
+63.3%
+60.9%
100%
100%
6
IWM Russell 2000
+61.4%
+56.0%
100%
67%
6
XLF Financials
+59.7%
+43.2%
100%
67%
6
XLI Industrials
+57.4%
+43.3%
100%
100%
6
The buy list for when the capitulation signal fires.
Historical pattern ideas, not financial advice. Past crash behavior varies by cause.
What each risk band was followed by
share of days
avg 12m return
P(12m drawdown >= 15%)
P(12m drawdown >= 25%)
band
Low <25
64%
11%
13%
2%
Moderate 25-45
19%
12%
20%
10%
Elevated 45-65
11%
3%
47%
30%
High 65+
5%
4%
63%
45%
Uses today's revised data, so live readings are noisier than this. Forward drawdown = worst S&P close over the next 252 trading days.