A brass bull figurine facing a brass bear figurine on a desk

Level of Bullishness: What Followed the S&P 500's Rarest Winning Streaks

Data through November 6, 2026 (S&P 500 closes from 1928, daily lows from 1962, VIX from 1990). Past market behavior doesn't guarantee future results.

The 2010 post at this URL measured the level of bullishness with price action: a long run of up closes and a long stretch without two down days in a row. Since 1928, the S&P 500 has gone 25 sessions without two lower closes in a row only 36 times. Three months after the 25th session, it was higher 61.1% of the time with a median gain of 1.63%, a little below the 64.0% and 2.57% of an average day.

In short, very strong runs were rare, and they didn’t lead to unusually strong follow-through. All figures use S&P 500 index data through November 6, 2026.

How rare are long winning streaks?

Six or more up closes in a row show up a couple of times a year on average (262 streaks in 99 years); ten or more has happened only twice since 1990.

Up closes in a row Streaks since 1928 Since 1990 Most recent (end date)
6+ 262 81 2026-06-02
7+ 135 39 2026-06-02
8+ 68 17 2026-06-02
9+ 32 7 2026-06-02
10+ 17 2 1995-09-14
11+ 9 2 1995-09-14
12+ 6 1 1995-09-14

The longest streak in the data is 14 sessions, ending April 15, 1971. The most recent long one was a 9-session run ending June 2, 2026. Up close means the index closed above the previous session’s close.

What followed N up closes in a row

Each row counts the day a streak reached that length, so every streak is counted once.

Set n 5d avg % 5d median % 5d % up 21d avg % 21d median % 21d % up 63d avg % 63d median % 63d % up
All sessions (baseline) 24830 0.16 0.30 56.22 0.66 1.05 60.27 2.01 2.57 64.04
7th straight up close 135 0.35 0.59 68.89 0.94 1.10 64.44 2.15 2.60 69.63
8th straight up close 68 0.08 0.20 60.29 0.82 1.02 63.24 2.64 3.27 72.06
9th straight up close 32 0.04 0.53 62.50 0.46 0.89 62.50 2.78 2.31 71.88
10th straight up close 17 0.28 0.67 76.47 0.62 0.51 70.59 1.65 1.56 58.82

On average, a long streak was not followed by a giveback. A week after the 7th straight up close, the index was higher 68.9% of the time, against 56.2% normally. By three months the rows sit close to the baseline. The 10-day row has 17 cases, too few to lean on.

25 sessions without two lower closes in a row

This is the 2010 post’s main measure: the index keeps closing higher, or never falls two days running, for at least 25 sessions. It shows up a few times a decade.

Decade Runs reaching 25 sessions
1920s 2
1930s 1
1940s 3
1950s 5
1960s 4
1970s 2
1980s 4
1990s 2
2000s 6
2010s 3
2020s 4

The longest such run lasted 40 sessions, ending June 13, 1950. The most recent run to reach 25 sessions did so on October 11, 2024. Returns from the 25th session:

Set n 5d avg % 5d median % 5d % up 21d avg % 21d median % 21d % up 63d avg % 63d median % 63d % up
1928-1961 signals 13 0.36 0.38 69.23 -0.56 -0.54 46.15 0.10 -1.37 46.15
1928-1961 all sessions 8508 0.13 0.32 55.85 0.58 0.93 58.13 1.79 2.36 59.59
1962-1989 signals 8 -0.41 -0.40 37.50 -0.36 -0.72 37.50 1.31 1.34 62.50
1962-1989 all sessions 7041 0.14 0.24 54.79 0.58 0.82 58.06 1.72 1.80 61.28
1990-2009 signals 8 0.11 0.31 62.50 0.43 0.62 62.50 3.24 3.75 62.50
1990-2009 all sessions 5043 0.14 0.24 55.60 0.61 0.94 61.02 1.84 2.51 66.21
2010-2026 signals 7 -0.45 -0.23 42.86 1.90 2.60 85.71 3.38 3.87 85.71
2010-2026 all sessions 4238 0.25 0.41 60.08 1.05 1.52 67.37 3.15 3.89 75.16
1928-2026 signals 36 -0.03 0.23 55.56 0.18 0.91 55.56 1.70 1.63 61.11
1928-2026 all sessions 24830 0.16 0.30 56.22 0.66 1.05 60.27 2.01 2.57 64.04

Over the whole history, the month after one of these runs averaged 0.18% against 0.66% for any session. Since 2010 the picture flipped: six of seven runs were followed by a higher index a month and three months later. Seven cases is a very small sample, and the eras disagree with each other.

The same test with daily lows

The 2010 post also counted sessions without two lower lows in a row. Daily lows exist in this data only from 1962.

Set n 5d avg % 5d median % 5d % up 21d avg % 21d median % 21d % up 63d avg % 63d median % 63d % up
All sessions 1962-2026 (baseline) 16322 0.17 0.29 56.41 0.71 1.08 61.38 2.13 2.65 66.38
Signals 1962-2026 19 -0.11 -0.11 42.11 0.25 1.24 63.16 0.92 2.25 57.89
Signals 1962-1989 9 -0.22 -0.59 44.44 -0.48 -0.05 44.44 0.87 2.06 55.56
Signals 1990-2026 10 -0.02 -0.09 40.00 0.90 1.90 80.00 0.96 3.06 60.00

The week after was weaker than usual (42.1% up against 56.4%), across only 19 runs. The most recent one reached 25 sessions on July 29, 2025.

What about sentiment surveys?

Many readers who search “level of bullishness” mean a survey, most often the AAII Investor Sentiment Survey. This page can’t test it because the history isn’t free. According to AAII’s survey page, the poll has run weekly since July 1987 and asks members whether they expect the stock market to be up, unchanged or down over the next six months. AAII says the full weekly history is available to members, and its spreadsheet download returned a bot-check page when we tried it on November 8, 2026, so we did not use it.

The closest free stand-in is the VIX, which reads calm or fear from S&P 500 option prices rather than from a poll. This table uses the first day in 21 sessions that the VIX closed at its lowest level of the prior 252 sessions (about a year):

Set n 5d avg % 5d median % 5d % up 21d avg % 21d median % 21d % up 63d avg % 63d median % 63d % up
All sessions (baseline) 9025 0.20 0.33 57.87 0.84 1.25 64.27 2.46 3.16 70.44
First VIX one-year low in 21 sessions 50 -0.28 -0.00 50.00 0.94 0.97 70.00 1.54 3.50 68.00
of which 1991-2009 27 -0.57 -0.28 44.44 0.17 0.66 59.26 0.86 1.98 74.07
of which 2010-2026 23 0.06 0.03 56.52 1.85 2.31 82.61 2.34 3.73 60.87

The week after a fresh VIX low was flat at the median, but the month after was up 70% of the time. Calm readings were not a dependable warning in either direction. The most recent one-year VIX low was 13.47 on December 24, 2025.

What these numbers do and don’t show

Strong price runs and calm VIX readings were uncommon, and what followed them stayed close to an ordinary day, with era-to-era swings bigger than the average effect. Signal counts run from 7 to 135, so many rows are small samples. Returns are S&P 500 price only (no dividends or costs), and Yahoo’s series has no Saturday sessions, which the NYSE still held in the early decades.

As of November 6, 2026, none of these readings was in play: the up-close streak stood at 1 session and the index had gone only 4 sessions since its last two lower closes in a row, after a 4.36% drop on November 5. For the fear side of that week, see what followed 20% VIX surges and the Williams’ VIX Fix. Another measure of a strong tape is the count of 52-week highs in December. All posts are on Market Stats.

Frequently asked questions

How often does the S&P 500 go up 10 days in a row?

Rarely. From 1928 through November 6, 2026, the S&P 500 had 17 streaks of 10 or more higher closes, and only 2 since 1990, the most recent ending September 14, 1995. The longest was 14 sessions, ending April 15, 1971.

Do long winning streaks lead to more gains?

Not by much. On the day a streak reached its 7th straight up close, the S&P 500 was higher five sessions later 68.9% of the time versus 56.2% for all sessions, but three-month results were close to average. After 25 sessions without two lower closes in a row, three-month returns were below average overall.

Where can I get AAII sentiment survey data?

AAII publishes the latest weekly reading on its Sentiment Survey page. It says the full weekly history back to 1987 is available to members. This page doesn't use it because the historical file is not freely downloadable.

Is a low VIX a sign of too much bullishness?

A low VIX means option prices imply calm, which is often read as complacency. Since 1991, the first VIX close at a one-year low in 21 sessions was followed by a flat median week (50% up, n = 50), but the next month was up 70% of the time, so it was not a reliable warning.

Sources

  1. S&P 500 index daily prices (^GSPC), Yahoo Finance (checked November 8, 2026)
  2. Cboe: VIX Index historical data (daily, 1990 to present) (checked November 8, 2026)
  3. AAII: Investor Sentiment Survey (method, history, member access) (checked November 8, 2026)

Jordan Whitfield

Editor, Trading the Odds

Jordan Whitfield edits Trading the Odds, which covers stock market statistics (seasonality, volatility and how markets behave around recurring events) and how prediction markets work and where they are legal. Market statistics on the site are calculated from public index data and say when the data runs through; every legal claim carries the date it was checked and a link to its source. How we check facts.