Every decision on your book depends on the shape of today's risk.
Before the market opens, our classifier tells you whether today's stress is real or fake, using data upstream of anything in your stack.
Volatility spiked while the system classified STABLE. The market recovered, confirming the call was correct.
March 2026. The system caught the crash and the recovery.
VIX caught neither.
The system left STABLE before the market dropped.
It classified SHIFTING. The S&P 500 was still at 6,817, with no drawdown underway and no conventional volatility alarm.
The drop the system flagged arrived.
The S&P 500 fell to 6,344, down about 7%. VIX spiked to 31, the alarm conventional tools wait for, three weeks after the system had already moved.
The system classified the episode over while VIX still flashed danger.
It classified RESOLVED with VIX still near 30. The S&P 500 then recovered 254 points within five trading days.
Iran Shock, March 2026. Live production classification. Historical research observation, not investment advice. Past performance does not guarantee future results.
When VIX spikes 25%, the system calls it: real or fake.
The call is delivered before the open, and a numbered Event Risk Brief is published the same day to track it until it resolves. The outcome is scored either way. Here is that call tested against every 25%+ VIX spike since 1990, all 123 of them.
WHIPSAW record, 1990–2025. Not investment advice; past performance does not guarantee future results.
A signal that moves the entire market. Nobody in finance reads it. We do.
Before the move
The cause, while it is still forming, before it ever reaches price.
- Space weather telemetry, upstream of price
- Independent of every signal in your stack
- Classifies stress across finance, space weather, and public health
After the move
The effect, reported once it has already shown up in price.
- VIX shows what traders already priced in
- Sentiment shows what was said after the move
- Price shows what traders already did
Why this isn't astrology.
Federal Reserve economists found the same correlation.
We discovered it in our own research. Krivelyova & Robotti (2003) reached the same finding, independently, in a Federal Reserve Bank of Atlanta working paper.
Decades of peer-reviewed science back the mechanism.
Cardiology, neuroscience, and environmental-health research all show space weather influencing human physiology and behavior.
Read the scientific foundation→It works in domains it was never fit to.
The same signal classifies stress across unrelated systems. Something that replicates that widely isn't curve-fit.
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.
This warning preceded real market trouble 64 of the 69 times it fired since 1990.
It fires about 2.0 times a year, on average 3.3 trading days before the trouble arrives.
Research Use Only: Not investment, financial, or trading advice. Past performance does not guarantee future results.
You get the classification at 07:30 ET. Here's the data.
Environmental and market signals are at baseline. No risk-shape escalation warranted.
Hedge premium decays here without doing protective work. The Positioning Edge captures the savings from half-sizing hedges during these stretches.
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 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.
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.
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.
VIX surged 25%+ over five sessions while the system held STABLE. The classifier is calling this a whipsaw: a volatility scare that will pass.
The false-scare filter. 64 of 67 correct since 2012 (95.5%). Of the 67 episodes, 41 scares faded on their own and 23 escalated but the system caught them in time. Three were misclassified.
The system returned to STABLE after a real escalation episode while VIX is still elevated above baseline. The crisis was real and now it is over.
The resolution call. The system says conditions have normalized even though conventional volatility signals still show stress. Canonical example: Iran 2026 (RESOLVED classified Mar 31 at VIX 25.10, SPX recovered +254 points over 5 sessions).
1990–2025 walk-forward validated (35 years). Full methodology at validation & methods.
Run the classifier against your own risk windows.
The system escalates before the drawdown. When the episode resolves, it returns to STABLE while volatility is still elevated. The pilot delivers that daily classification every trading morning for 60 days, with the historical archive and API access to replay every call against your own book.
60-Day Evaluation Pilot
- Daily pre-market classification for 60 days
- Historical classification archive (2020-present)
- API access
- Event Risk Briefs
Credentials and the archive arrive by email. Your first live classification arrives the next trading morning at 07:30 ET.
Not investment advice. Past performance does not guarantee future results. Mindforge is not a registered investment adviser. Full terms at mindforge.tech/terms.
The same architecture, rebuilt for seven more markets.
When a market's volatility spiked while the classifier was still calling STABLE, there were two ways for it to be right: the spike could fade on its own, or the classifier could move off STABLE before the market fell further. Each bar is how often one of those two outcomes happened.
Backtested 2000–2024 (MSCI EM from 2003), ablation-validated, not yet in live production. 192–302 spike events per market. Exploratory research; past backtested performance does not indicate future results.