When it happens · method, not a ticket

Buy the washout?

The street line is “buy oversold, buy high VIX.” The ended tapes on this desk say something colder: oversold is a wash, high VIX is insurance being bid, and the print that paid was three months after a trough named in hindsight. Confirm two clocks. Do not catch the first knife. A peak is not called. A trough is not called.

Live stage

Quiet

Quiet stage. From the 52-week high the tape is orderly, the VIX slope is ordinary, labor is quiet. Quiet is not a buy ticket. If a winter comes, the median trough photograph is +18.9% in a quarter — hindsight. A peak is not called.

Quiet
Watch
Stress
Crash

S&P RSI-14

49.0

Neutral

VIX 3m − spot

1.6

contango

HY OAS

3.30%

tight

From 52-week high

-2.8%

2026-10-06

Photograph 01 · RSI

Oversold is a wash

RSI 49.0 · Neutral

Washed-out on this desk is RSI ≤ 30 and at least 5% below the 50-day. 2020 printed that. So did the 2022 autumn. The first wash is often not the low. Averaging at RSI 32 is how 1987 and 2020 punished. This is not a buy ticket.

Photograph 02 · VIX

High VIX is insurance bid

VIX 18.1 · +1.6 contango

Backwardation is the front being bid. Crisis on this desk is spread ≤ −5. 16 Mar 2020 is the anatomy snapshot on the VIX board. A high VIX is not a timer that the low is in. It is the cost of insurance going up. Confirm a second sleeve.

Photograph 03 · Credit

OAS wide is the fuse

3.30% · tight

ICE BofA US High Yield OAS. Wide is the credit photograph next to the hyperscaler still. 2008 and 2020 went through double-digit OAS. Tight OAS into a rich CAPE is the other winter. Neither is a ticket. Credit is a second clock.

Photograph 04 · Aftermath

The trough that paid

Median from trough, 3m +18.9% · from high -8.8%

From a known high the median S&P photograph is still negative at three months. From a trough named in hindsight it is the one that paid. Fastest −20% on this tape: 2020 in 22 days. Re-entry is a rule after a higher low, not a call that this is the low.

Rules from ended tapes

What the winters actually did

01
Confirm more than one clock
A −5% print is not a crash. Wait for a second sleeve: VIX slope, NFCI, junk OAS, or Sahm. 1987 and 2020 went through −20% in days. 1973 and 2000 took months. Speed is the tell, not the first red week.
02
Ballast, do not forecast
From a known high the median S&P photograph is still negative at three months. Cash and bills are ballast. 2022 is the warning that long bonds do not always hedge. Gold exists on the modern winters only. This is not a ticket.
03
Do not catch the first knife
Days-to-−10 and days-to-−20 sit on the speed board. Averaging at −5% is how 1987 and 2020 punished. Size down if Stress lights. Do not all-out sell on one sleeve. Do not all-out buy a bounce you cannot name as a trough.
04
Trough is hindsight
The five-session bounce from a trough is violent and easy to miss. The three-month trough photograph is the one that paid. Re-entry is a rule after a higher low, not a call that this is the low. A peak is not called. A trough is not called.
05
Scale in. Do not lever.
Month-end overlapping photographs after −10/−20/−30 are green at three and five years — that is being invested while down. The first print is colder: most first −10% crossings went further, and every first −30% print in the sample did. A slice at −10, dry powder at −20, no loan at −30. Waiting for −30% misses the ordinary corrections that never become bears. A peak is not called.

S&P 500 · Shiller total return · 1950–now · recomputed

Forward returns after corrections

Overlapping month-end total returns since 1950 — the same room as Carlson’s March 2026 card. Being invested while down, not a ticket on the first red print. A peak is not called.

Live analog is -2.8% from the 52-week high · not a correction yet. The photograph is what paid after the print, not a call that one is coming.

Average total return · overlapping months

−10% went further

65%

first print, next 3y

−30% went further

100%

every first −30% in sample

First −10% count

independent crossings

Overlapping −10%

294

months in the hole

Four rules from the photograph

01
The green card is overlapping months
Carlson’s photograph counts every month the S&P finished −10/−20/−30% from its high, then the next 1 / 3 / 5 years of total return. 2008–09 sits in that sample many times. That is “being invested while down,” not a single ticket on the first red print. This desk recomputes that room live.
02
The first print still has a knife
Independent crossings — first month through −10%, then wait for a new high — are rarer and colder. Most first −10% prints went further before the low. Every first −30% print in the sample did. Averaging the whole book at the first −10% is how 2000 and 2008 punished.
03
Scale in. Do not lever.
A slice at −10, more at −20, dry powder at −30. Waiting for −30% misses the ordinary corrections that never become bears — most −10% prints do not. A loan at −30% is how a further −20% becomes a ruin. Horizon is years. One year still loses.
04
Confirm a second clock
A −10% at CAPE 35 with tight junk OAS is not 2009. Credit, VIX slope, NFCI, Sahm sit on this desk for that reason. Stay invested beats selling the dip and buying the new high. Past photographs do not guarantee the next winter. A peak is not called. A trough is not called.

First −10% prints · 17 crossings

One row per independent crossing. Further is extra price decline in the next 36 months after you would have bought. Recover is months back through the buy print, price only. CAPE is Shiller PE10 that month.

Print	Depth	CAPE	+1y	+3y	+5y	Further	Recover
2025-04	-11.1%	—	+30.9%	—	—	0.0%	1m
2022-05	-13.6%	30.7	+4.4%	+50.2%	—	-7.8%	3m
2020-03	-19.1%	24.8	+50.0%	+56.9%	+130.9%	0.0%	1m
2018-12	-11.5%	28.3	+26.2%	+91.7%	+98.3%	0.0%	1m
2008-01	-10.5%	24.0	-35.6%	-0.2%	+20.0%	-45.1%	4m
2000-12	-10.4%	37.3	-12.8%	-14.9%	+2.8%	-37.1%	1m
1998-09	-11.8%	33.5	+30.8%	+6.3%	+7.2%	0.0%	1m
1990-09	-12.5%	15.3	+27.0%	+59.7%	+112.4%	-2.6%	3m
1987-10	-14.9%	15.5	+2.5%	+21.3%	+73.7%	-14.0%	15m
1981-09	-12.8%	7.6	+9.6%	+62.8%	+152.7%	-7.5%	1m
1973-06	-11.5%	15.8	-11.3%	+9.6%	+15.0%	-36.0%	1m
1969-07	-11.1%	18.7	-17.1%	+25.1%	-1.3%	-20.2%	3m
1966-08	-13.6%	19.9	+21.1%	+28.4%	+42.2%	-4.4%	3m
1962-05	-12.2%	19.1	+15.2%	+55.7%	+72.2%	-11.7%	8m
1960-10	-10.1%	16.6	+30.5%	+49.5%	+98.5%	0.0%	1m
1957-02	-10.9%	15.8	-1.2%	+42.9%	+91.7%	-7.2%	1m
1953-09	-11.1%	11.1	+42.6%	+129.8%	+159.9%	0.0%	1m

Recomputed, not copied

Carlson (A Wealth of Common Sense, 29 Mar 2026) took every month-end the S&P was down 10%, 20%, or 30% from its high, then the forward 1 / 3 / 5-year total return. His card printed −10% +15 / +42 / +72, −20% +17 / +45 / +74, −30% +21 / +48 / +88. This desk rebuilds that overlapping room from Shiller monthly prices with trailing dividends reinvested, 1950 through the last complete month. Means sit next to his card on each tile. Medians and win rates sit under them. First-print is the independent-crossing cousin — the decision you actually make the first time the tape tags −10%. Failed corrections are first −10% prints whose 36-month peak-to-trough never reaches −20% — the ordinary dips that never become bears. Waiting for −30% misses them.

Overlapping months overweight winters that lasted (1973–74, 2000–02, 2008–09). That is why the viral card looks greener than the first print, and why a loan at −30% is the inverted strategy. 2008’s first −30% month went another twenty points before the low. Stay invested and scale in. Do not catch the first knife. Past performance does not guarantee future results. When it happens · Method.

Live clocks

Confirm more than one

From 52-week high

-2.8%

First −5% is a watch, not a call

VIX slope

+1.6 · contango

Backwardation is insurance being bid. Crisis is ≤ −5.

NFCI

-0.46 · loose

Tight is a credit photograph, not a sell ticket.

Sahm

0.00 · quiet

Triggers at 0.50. A recession clock, not a peak.

RSI tape

49.0 · neutral

Overbought is fragility. Washed-out is the other side.

Junk OAS

3.30% · tight

Wide spreads are the credit fuse next to the hyperscaler still.

Equal-weight

-24.8%

RSP/SPY from its 52-week high. Narrow tape is a photograph.

Buffett indicator

230%

Market cap / GDP. A price, not a crash clock.

Shiller CAPE

30.8

Ten years of real earnings. Mean reversion is the photograph.

36-month Pearson · monthly last · country ETFs

Gold, bitcoin, and the world tape

36-month S&P/gold correlation is 0.07. S&P/bitcoin is 0.48 — a risk sleeve, not ballast. 2020 gold held while bitcoin sold with the tape, then ripped. 2022 both rhymed with Nasdaq. Haven is a photograph after the winter, not a ticket. Country tapes are iShares ETFs. A peak is not called.

S&P / gold

-0.2

0.07

36-month

S&P / bitcoin

-0.1

0.49

risk sleeve

Gold, 1y

-8.3%

GLD

Bitcoin, 1y

-20.9%

BTC-USD

Window · S&P vs gold and bitcoin · 10Y · 2016-11-12 → 2026-11-12

Months from known high · 2020 and 2022 · indexed 100

Dashed is bitcoin. Gold held in 2020; both sold with Nasdaq in 2022. Haven is not automatic.

Any two tapes

Pick a pair. Scatter is monthly log returns, last 120 months. Rolling is 36-month Pearson.

0.49now

Window · selected pair · 10Y · 2016-11-12 → 2026-11-12

Monthly log returns · percent

	S&P	Japan	Germany	United Kingdom	China	Brazil	Emerging	Nasdaq 100	Russell	Tech	High yield	Long bonds	Gold	Bitcoin
S&P	1.00	0.47	0.40	0.25	0.10	0.09	0.60	0.88	0.79	0.80	0.66	0.33	0.07	0.47
Japan	0.47	1.00	0.70	0.65	0.18	0.49	0.47	0.48	0.28	0.44	0.30	0.20	0.27	0.33
Germany	0.40	0.70	1.00	0.81	0.38	0.39	0.36	0.27	0.29	0.19	0.48	0.26	0.37	0.29
United Kingdom	0.25	0.65	0.81	1.00	0.35	0.43	0.24	0.04	0.30	-0.02	0.44	0.33	0.35	0.07
China	0.10	0.18	0.38	0.35	1.00	0.26	0.21	-0.02	-0.04	-0.06	0.23	0.02	0.24	0.21
Brazil	0.09	0.49	0.39	0.43	0.26	1.00	0.38	0.03	0.18	-0.05	0.21	0.06	0.49	-0.05
Emerging	0.60	0.47	0.36	0.24	0.21	0.38	1.00	0.67	0.49	0.65	0.38	0.29	0.50	0.09
Nasdaq 100	0.88	0.48	0.27	0.04	-0.02	0.03	0.67	1.00	0.60	0.97	0.41	0.15	-0.02	0.46
Russell	0.79	0.28	0.29	0.30	-0.04	0.18	0.49	0.60	1.00	0.50	0.75	0.52	0.09	0.28
Tech	0.80	0.44	0.19	-0.02	-0.06	-0.05	0.65	0.97	0.50	1.00	0.30	0.12	-0.08	0.43
High yield	0.66	0.30	0.48	0.44	0.23	0.21	0.38	0.41	0.75	0.30	1.00	0.72	0.26	0.16
Long bonds	0.33	0.20	0.26	0.33	0.02	0.06	0.29	0.15	0.52	0.12	0.72	1.00	0.20	-0.11
Gold	0.07	0.27	0.37	0.35	0.24	0.49	0.50	-0.02	0.09	-0.08	0.26	0.20	1.00	-0.05
Bitcoin	0.47	0.33	0.29	0.07	0.21	-0.05	0.09	0.46	0.28	0.43	0.16	-0.11	-0.05	1.00

S&P

SPY

758

+10.9%

Japan

EWJ

91.68

+10.4%

Germany

EWG

42.85

+4.7%

United Kingdom

EWU

47.77

+10.7%

China

FXI

33.56

-15.1%

Brazil

EWZ

31.69

-5.7%

Emerging

EEM

61.06

+12.4%

Nasdaq 100

QQQ

688

+11.2%

Russell

IWM

279

+12.3%

Tech

XLK

174

+21.3%

High yield

HYG

77.07

-4.8%

Long bonds

TLT

79.90

-11.4%

Gold

GLD

356

-8.3%

Bitcoin

BTC-USD

71,491

-20.9%

Social tape · VIX zones · Best practices from the masters · Back to the desk