Volatility spiked while the system classified STABLE. The market recovered, confirming the call was correct.
Every decision on your book depends on the shape of today's risk.
Before the market opens, our classifier tells you which of five risk environments the day is setting up.
When volatility spiked but it said STABLE, it was right 95% of the time, because it's reading what nothing else is reading.
35 years validated. 12/13 on CRISIS. 95% on STABLE while VIX elevated.
The single miss: GameStop, January 2021. The system flagged the crisis — VIX hit 37 — but the spike reversed within the scoring window. Full case study
One regime shift. Two decisions. Both before the market moved.
SPX + VIX vs. MRC State Classification
March 4, 2026. The system classified SHIFTING. VIX was 19. The S&P 500 was at 6,853. No traditional volatility signal was elevated.
Over the following three weeks, VIX rose to 31 and the S&P 500 fell 7.4% to 6,344. A desk monitoring the daily classification saw the regime shift before any VIX-based trigger fired.
March 31. The system classified STABLE. VIX was still at 25, elevated by conventional standards. The S&P 500 recovered 254 points within five trading days.
Iran Shock, March 2026. Out-of-sample event. Historical research observation, not investment advice. Past performance does not guarantee future results.
You get the classification at 07:30 ET. Here's the data.
| Classification | Frequency | Precision | Avg lead to hit | VIX at classification | SPX median 5d | VIX median 5d |
|---|---|---|---|---|---|---|
| STABLE▼ | ~89% of days | — | — | 17.6 | +0.3% | ~0% |
Environmental and market signals are at baseline. No risk-shape escalation warranted. Hedge premium decays here without doing protective work. The Stability Dividend captures the savings from half-sizing hedges during these stretches. | ||||||
| SHIFTING▼ | ~2.0/yr | 93% (64/69) | 3.3d | 23.1 | +0.2% | — |
Conditions are shifting away from STABLE. The environmental signal has moved before price has confirmed. The earliest warning. VIX is often still in the teens when SHIFTING fires. Median 2-day lead before stress appears in price-derived indicators. The trigger to restore full hedge sizing. | ||||||
| ELEVATED▼ | ~1.4/yr | 76% (37/49) | 1.6d | 23.5 | +1.0% | — |
Elevated risk conditions detected. Environmental gate plus VIX floor confirms the regime is not transient. A confirmation tier. Roughly 1 in 4 ELEVATED classifications does not produce the expected forward stress. Pair with desk-level evidence before sizing aggressively. | ||||||
| SHOCK▼ | ~1.9/yr | 85% (57/67) | 0.0d | 26.3 | +1.3% | — |
Rapid short-window volatility expansion. VIX rate-of-change exceeds the acute-path threshold. VIX is already elevated at classification. The forward path typically shows mean reversion: the spike is already in progress, and the median outcome is recovery. | ||||||
| CRISIS▼ | ~0.4/yr | 92% (12/13) | 1.9d | 32.2 | -0.9% | — |
Systemic financial-system stress. The most severe classification the system issues. The forward path typically shows further deterioration before recovery. The April 2025 tariff crisis was the first live CRISIS classification; VIX touched 60 intraday the next session. | ||||||
| Stable(Watch) | ~5.2/yr | 96% (64/67) | — | 18.9 | +1.9% | -13.4% |
1990–2024 walk-forward validated. Full methodology at validation & methods.
A signal that moves the entire market. Nobody in finance reads it. We do.
VIX, sentiment, options flow, price patterns. They all tell you what already happened.
- ✗ VIX tells you what traders already priced in
- ✗ Sentiment tells you what people said after the market moved
- ✗ Price patterns show you what traders already did
Solar and geomagnetic data that shapes human decision-making before it shows up in price.
- ✓ Solar and geomagnetic data reads the cause of market behavior, not the effect
- ✓ This data is independent of every signal in your existing stack
- ✓ The same signal classifies stress in finance, space weather, and public health independently
Why this isn't astrology.
We discovered the correlations through our own research. Federal Reserve economists (Krivelyova & Robotti, 2003) independently documented the same market correlation in a Federal Reserve Bank of Atlanta working paper.
Decades of peer-reviewed research across cardiology, neuroscience, and environmental health show that solar and geomagnetic activity influence human physiology and behavior.
Read the scientific foundation→The same signal classifies stress in systems it was never fit to. A signal that replicates across unrelated domains isn't curve-fit.
The same classifier, tested blind across seven global markets.
When the classifier said STABLE and the market's volatility spiked anyway, the classifier was right 72–90% of the time across all seven markets.
| Market | Episodes | Precision |
|---|---|---|
| Hang SengHK / China | 23 | 96% |
| Luxury BasketEuropean Luxury Equities | 20 | 85% |
| MSCI EMEmerging Markets | 27 | 81% |
| Nikkei 225Japan | 65 | 75% |
| FTSE 100United Kingdom | 46 | 74% |
| ASX 200Australia | 19 | 68% |
| DAXGermany | 48 | 67% |
ELEVATED tier precision shown. Backtested 2000–2024.
Your strategy doesn't change. Your sizing does.
MRC classifies the market into one of five risk shapes before the open. You feed the classification into the trades you were already going to make. No strategy change. No new infrastructure. The classification sits upstream of your stack.
Backtested 1990-2025. Numbers from canonical manifests + walk-forward.
87% of days are STABLE. Hedging through them costs premium.
Most funds hold static hedges through long STABLE regimes, paying premium drag the whole time. A daily STABLE classification justifies reduced hedge sizing when conditions are confirmed STABLE, then re-size when the regime shifts.
illustrative backtest · simple notional-weighting policy
One classification. Three ways to receive it.
Widened distribution. VIX median 23.1. The S&P 500 moved a median +0.2% over the next 5 sessions. The market could move sharply or settle back. Most of the time, conditions returned to STABLE.
Research Use Only: Not investment, financial, or trading advice. Past performance does not guarantee future results.
Start with a pilot. Stay for the edge.
60-Day Pilot
- Daily pre-market classification for 60 days
- Full historical dataset (1990-present)
- API access
- Anomaly Reports
- 100% credited against annual if you continue
Annual
- Everything in the pilot
- Continuous daily classifications
- Priority support
- Founding cohort rate while active
Not investment advice. Past performance does not guarantee future results. Mindforge is not a registered investment adviser. Full terms at mindforge.tech/terms.