With the S&P 500 reaching a new all-time high in early August 2026, a pronounced contrarian bearish divergence has emerged among market participants. Retail "Dumb Money" Confidence (red line in the chart below) has surged from neutral into optimistic territory at 0.61, while institutional "Smart Money" Confidence (blue line) has dropped into pessimistic territory at 0.31.
Historically, the most dangerous periods for stocks are when dumb money is high and smart money is low at the same time the index is elevated. It does not mean an immediate crash is guaranteed, but it raises the odds of a meaningful pullback or at least a period of weaker returns ahead.
.png)
projection points to a mid-October major low.
Static projection based on the latest 20-week cycle period (17.29 weeks / 121 CD).
However, having likely printed a nominal 20-week Hurst cycle low on June 30 (17.29 weeks / 121 CD off the late-March 40-week trough), the S&P remains in the rising phase of its second 20-week sub-cycle—part of a broader 40-week (9-month) and 18-month structure targeting a major low in late November (± 8 weeks).
The current rise suggests to extend into early September—interrupted by the 5- and 10-day cycle higher lows of August 5–6 (Wed-Thu) and August 11–12 (Tue-Wed) and a 20-day cycle low around August 24 -28 (Mon-fri)—before yielding a late-September 80-day (10-week) cycle trough. A secondary rally into mid-October—capped below the September peak—is then likely to trigger the final decline into the primary 40-week and 18-month cycle trough in November.
Average S&P 500 total-return path (indexed to 100 on midterm Election Day) for all midterm years since 1970 (1970–2022), spanning roughly ±6 months. X-axis centers on Election Day (first Tuesday in November); y-axis tracks cumulative total return. The average line rises in the final ~22 trading days before the election (= October 2, 2026) and continues higher afterward (+14.1% average in the following six months). A separate “Lost Control” series (party loses presidential trifecta) lags the broader average post-election (+10.4% vs. +16.1%).
