Dashboard › Research › Market Regime Backtest
Market Regime Backtest
What the S&P 500 actually did over the 4 and 13 weeks after each daily regime call, from 2026-07-09 to 2026-11-21 — 266 calls in total.
Results by regime
| Regime | Calls | Next 4 weeks: median | Next 13 weeks: median |
|---|---|---|---|
| Expansion | 152 | +0.75% (66% positive, n=150) | +2.71% (65.5% positive, n=84) |
| Transition | 108 | -1.65% (25.4% positive, n=71) | +3.96% (85.7% positive, n=7) |
| Unknown | 6 | +0.58% (83.3% positive, n=6) | -3.73% (25% positive, n=4) |
How it is measured
- Every daily run logs its regime call. For each call, the forward S&P 500 price return over the next 4 and 13 weeks is computed from index closing prices.
- “Positive” is the share of calls followed by a positive return over that horizon; the median is the middle forward return.
- Calls are made daily, so forward windows overlap heavily. The effective number of independent observations is far smaller than the call count.
Read this before using the numbers
- The history is short. With fewer than 30 observations per regime, the results show direction only; they are not statistically meaningful.
- The sample is unbalanced. Most calls so far fall in one regime, so the other rows rest on a handful of observations.
- Returns are price-only, before costs and taxes, and are not a simulation of any portfolio.
- This is a record of what happened after the calls, not evidence that the calls caused or predicted the outcome.
How the regimes are defined: Market Regime Indicator · Data and limits: Methodology