Market breadth dashboard · Long-history scenarios
Frozen data through 2026-10-02. Current onset: 2026-09-18; current episode is on session 11.
Exploratory study using the current 504-stock survivor cohort with the project's fixed denominator. Missing historical prices and membership limitations affect early episodes. These are historical frequencies, not calibrated forecasts. The ongoing episode is excluded.
Exclude all segments in a completed chain of at least three pink episodes, separated by at most five non-pink sessions, whose total first-onset-to-final-exit span is at most 40 trading sessions. Retained episodes keep their original dates, lows and returns. No episodes are merged in this report.
Excluded onset dates: 2015-03-11, 2015-03-26, 2015-04-17. Among near-high onsets, excluded: 2015-04-17.
For this snapshot the near-high cohort falls from 15 to 14 episodes. The removed near-high episode lasted only one session, so the nine session-11 cases and four session-11 near-high cases are unchanged.
This is a retrospective noise-filter sensitivity study. Whether a full rapid-switch chain forms is known only afterward; the exclusion is not a signal available on the initial onset or necessarily on session 11. It does not establish a more accurate live forecast.
2018-10-05 remains in the onset and session-11 cohorts. It falls outside the narrower session-11 near-high cohort because its price was already 5.48% below its trailing high.
Near a high means the onset close is no more than 2% below the highest adjusted SPY close in the trailing 63 trading sessions, including onset. This 63-session lookback defines the prior high; outcomes follow the entire pink episode, without a 63-session cutoff.
The current onset was -1.84% below this high. The latest close is -0.81% below its trailing high. A trailing 252-session high selects 14 retained completed onset episodes at the 2% threshold in this dataset.
| Cohort | n | More than 5% | More than 10% | More than 20% |
|---|---|---|---|---|
| All high-price onsets | 14 | 5/14 (35.7%) | 4/14 (28.6%) | 0/14 (0.0%) |
| Still pink on session 11 | 9 | 5/9 (55.6%) | 4/9 (44.4%) | 0/9 (0.0%) |
| Still pink AND within 2% of high on session 11 | 4 | 2/4 (50.0%) | 1/4 (25.0%) | 0/4 (0.0%) |
Declines compare the first minimum close during pink with the onset close. Thresholds are strict. Counts overlap. Conditioning on session 11 matches today's observed pink duration; the additional high-price condition is evaluated at that same historical anchor, without using future prices.
| Metric (returns: %, times: sessions) | Mean | Median |
|---|---|---|
| Onset to price low: return (%) | -8.27 | -6.03 |
| Onset to price low: trading sessions | 38.44 | 24.00 |
| Price low to first non-pink day: trading sessions | 31.44 | 25.00 |
| Onset to first non-pink day: trading sessions | 69.89 | 46.00 |
| Onset to first non-pink day: return (%) | 0.43 | 1.53 |
From the session-11 anchor, 60 sessions later the price was lower in 4/9 cases (44.4%). This uses the anchor price, whereas the full-episode decline table uses the onset price.
| Onset | Gap to high | Pink at session 11 | Price-low date | Low return | Sessions to low | Sessions to end | End return |
|---|---|---|---|---|---|---|---|
| 1996-10-30 | -1.48% | No | 1996-10-30 | 0.0% | 0 | 1 | +1.0% |
| 1998-05-19 | -1.55% | Yes | 1998-08-31 | -13.5% | 72 | 131 | +7.9% |
| 2005-06-10 | -0.47% | No | 2005-06-10 | 0.0% | 0 | 4 | +1.0% |
| 2005-08-19 | -1.81% | Yes | 2005-08-26 | -1.4% | 5 | 15 | +1.5% |
| 2005-09-14 | -1.22% | Yes | 2005-10-13 | -4.3% | 21 | 46 | +1.6% |
| 2006-04-17 | -1.79% | No | 2006-04-17 | 0.0% | 0 | 1 | +1.6% |
| 2007-10-16 | -1.73% | Yes | 2008-03-10 | -16.3% | 99 | 135 | -9.2% |
| 2014-04-25 | -1.37% | Yes | 2014-04-25 | 0.0% | 0 | 21 | +2.8% |
| 2014-07-25 | -0.46% | Yes | 2014-08-07 | -3.4% | 9 | 24 | +1.2% |
| 2014-09-12 | -0.98% | Yes | 2014-10-16 | -6.0% | 24 | 41 | +2.9% |
| 2015-05-05 | -1.30% | No | 2015-05-06 | -0.4% | 1 | 10 | +2.0% |
| 2015-05-28 | -0.49% | Yes | 2015-08-25 | -11.4% | 62 | 129 | -0.8% |
| 2016-12-01 | -0.88% | No | 2016-12-01 | 0.0% | 0 | 6 | +3.2% |
| 2018-10-05 | -1.52% | Yes | 2018-12-24 | -18.1% | 54 | 87 | -4.2% |
The four cases still near a high on session 11 began 2005-08-19, 2014-04-25, 2014-09-12 and 2015-05-28. Only four observations remain: absence of a 20% decline here does not establish zero future risk.
| Onset gap allowed | n | More than 10% decline |
|---|---|---|
| 1% | 3 | 1/3 |
| 2% | 9 | 4/9 |
| 5% | 21 | 8/21 |
Near-high filtering reduces sample size and may leave dependent episodes from the same market period. No entry rule or causal conclusion is validated by this subset.