Market top risk today

Is weakness building beneath strong prices? Six warning signs that have tended to appear before market tops, combined into one score from 0 to 100.

Top risk, Oct 16, 202645 Watch

SPY below its 1-year high2.0% the score fades to zero 5% below the high

The weights are a published design, not fitted to past declines, and the score is not a probability of a fall. High means more warning evidence, not a forecast. In a market already well below its high the score reads low: that is weakness already here, which the health score measures.

Two different lenses on market risk

Top risk asks whether several warning signs that have come before market tops are building up at once: six components, weighted into one score. Divergence asks whether the broad market is confirming the index’s strength: five breadth families either confirm SPY’s new highs or fail to. They complement each other, and neither predicts an exact turning point.

What makes up the score today

The six components today
Component	Weight	Reading (0-100)	Points in the score
Advance-decline divergence	25%	59	14.8
New-high divergence	20%	50	10.0
50-day participation falling	20%	59	11.9
Equal weight behind SPY	15%	6	0.8
Up/down volume weak	10%	13	1.3
Fewer sectors with new highs	10%	59	5.9
Top risk score	100%		44.8

How each part is measured

Advance-decline divergence (25%): how far the advance-decline line is below its own 63-day high, in average days of net advancing and declining stocks; full marks at 3 days.
New-high divergence (20%): how far the share of stocks at 52-week highs (10-day average) has fallen from its 63-day peak; full marks at 60% below.
50-day participation (20%): how many points the share of stocks above their 50-day average is below its 63-day peak; full marks at 30 points.
Equal weight against SPY (15%): how far the 500 largest companies, equal amounts in each, trail SPY over 63 days; full marks at 10 points behind.
Up/down volume (10%): how far the 10-day ratio of dollar volume in rising stocks to falling stocks is below 1; full marks at 0.5.
Sector leadership (10%): how many fewer sectors have fresh 52-week highs than at their 63-day peak; full marks at 4 fewer.
Each part is scaled by how close SPY is to its 1-year high: full weight at the high, none 5% below it.
Bands: 0-24 Low, 25-49 Watch, 50-74 Elevated (several warning families agree), 75-100 High.

The score since Mar 19, 2018

Top: SPY, shaded where the score was Elevated or High (50 or more). Bottom: the score.

SPY
Elevated or High
Top risk score

How it has done

Over 8.6 years the score was Elevated or High on 5% of days, in 22 separate spells; 1.5 a year were not followed by a 5% fall within 60 trading days. It reached Elevated in the 3 months before 2 of the 4 falls of 10% or more, a median 27 trading days before the high. SPY-or-T-bills, stepping aside at 50 or more, returned 13.5% a year against 14.6% for SPY.

SPY after each band

SPY after each band
Band	Days	Separate spells	SPY next 3 months (avg)	Worst fall in the next 3 months (avg)	Followed by a 10% fall within 3 months
Low	1,611	70	+4.6%	−4.6%	14% (220 of 1586 days)
Watch	440	92	+1.7%	−5.7%	15% (63 of 421 days)
Elevated	100	27	+1.3%	−5.8%	25% (21 of 85 days)
High	7	5	−6.8%	−11.2%	100% (3 of 3 days)
All days	2,158	–	+3.9%	−4.9%	15% (307 of 2095 days)

Counted in trading days. Neighbouring days overlap and are not independent: the 7 High days came from 5 separate spells, so a figure such as 100% rests on very few real events. Historical results do not establish future reliability.

Warnings followed by a fall

Elevated spells: followed by a fall
Elevated spells followed by	within 30 days	within 60 days	within 90 days
A fall of 5% or more	6 of 21	8 of 21	8 of 21
A fall of 10% or more	1 of 21	4 of 21	6 of 21
A fall of 15% or more	0 of 21	2 of 21	4 of 21

Every decline of 10% or more

SPY declines of 10% or more and whether a warning came first
High	Low	Fall	Warning in the 3 months before the high	How long before
Sep 20, 2018	Dec 24, 2018	−19.3%	No	–
Feb 19, 2020	Mar 23, 2020	−33.7%	No	–
Jan 3, 2022	Oct 12, 2022	−24.5%	Yes	18 days
Feb 19, 2025	Apr 8, 2025	−18.8%	Yes	36 days

Every warning (22), newest first

Every warning and SPY afterwards
Started	Lasted (days)	SPY lowest within 30 days	60 days	90 days	Within 60 days
Sep 21, 2026	19	–	–	–	–
Aug 8, 2025	2	−0.3%	−0.3%	−0.3%	False alarm (no 5% fall)
Jan 22, 2025	1	−5.6%	−17.9%	−17.9%	Fall of 10%+
Dec 24, 2024	2	−3.5%	−8.3%	−17.2%	Fall of 5-10%
Jun 25, 2024	11	−5.0%	−5.0%	−5.0%	Fall of 5-10%
Jun 13, 2024	6	−0.4%	−4.3%	−4.3%	False alarm (no 5% fall)
Jun 11, 2024	1	0.6%	−3.3%	−3.3%	False alarm (no 5% fall)
Feb 5, 2024	3	0.3%	0.3%	0.3%	False alarm (no 5% fall)
Jan 16, 2024	4	−0.6%	−0.6%	−0.6%	False alarm (no 5% fall)
Jan 12, 2022	1	−10.4%	−11.6%	−17.1%	Fall of 10%+
Dec 21, 2021	10	−6.9%	−10.1%	−10.8%	Fall of 10%+
Dec 7, 2021	8	−4.3%	−9.6%	−10.8%	Fall of 5-10%
Aug 19, 2021	9	−2.1%	−2.2%	−2.2%	False alarm (no 5% fall)
Aug 16, 2021	2	−2.7%	−3.8%	−3.8%	False alarm (no 5% fall)
Aug 3, 2021	5	−0.5%	−2.5%	−2.5%	False alarm (no 5% fall)
Jul 20, 2021	9	0.8%	−0.2%	−0.2%	False alarm (no 5% fall)
Jul 14, 2021	2	−2.6%	−2.6%	−2.6%	False alarm (no 5% fall)
Apr 13, 2021	1	−1.8%	−1.8%	−1.8%	False alarm (no 5% fall)
Apr 7, 2021	1	−0.3%	−0.3%	−0.3%	False alarm (no 5% fall)
Mar 26, 2021	4	−0.3%	−0.3%	−0.3%	False alarm (no 5% fall)
Aug 26, 2020	1	−6.8%	−6.8%	−6.8%	Fall of 5-10%
Oct 2, 2018	5	−9.5%	−19.1%	−19.1%	Fall of 10%+

Live record

Live since the close of Oct 16, 2026: every reading after that day is kept exactly as the daily run made it (0 so far). Version 1.

A small number of real declines since 2017 means any of these figures can be luck. The weights are not tuned to them.

Free · every Friday after the close

The health score and the top-risk score, and how far each moved
Every indicator that turned bullish or bearish, and why
Whether breadth confirmed the S&P 500’s new highs
The charts that matter, in under two minutes

See the latest report (issue of Oct 16, 2026) →