Skip to main content
Mindforge LogoMINDFORGE INTELLIGENCE
Mindforge Intelligence

Risk Shapes Reference

Seven named risk shapes. Each with defined activation logic and validated precision (19902025). Not investment advice.


1. Overview

The Market Risk Classifier assigns a risk shape to every US trading day. Each risk shape has defined trigger conditions and a scored precision record.

Risk ShapeActivationsHitsFPPrecisionFrequency
CRISIS1312192%~0.4/yr
SHOCK67571085%~1.9/yr
ELEVATED49371276%~1.4/yr
SHIFTING6964593%~2.0/yr
STABLEBaseline classification: ~87% of trading days
WHIPSAW123117695%~3.5/yr
RESOLVED71521973%~2/yr

Historical backtested performance (1990-2025). Precision = hits / activations. BD = NYSE/Nasdaq trading days.

Historical backtested performance. Live operation began 2025. Past classifications do not guarantee future accuracy. Full scoring methodology at mindforge.tech/validation-and-methods.

2. Risk Shapes

What each risk shape means, when it activates, and how institutional teams use it. Severity order: CRISIS > SHOCK > ELEVATED > SHIFTING > STABLE.

2.1 CRISIS

Crisis-grade classification for sustained extreme market dislocations. Ultra-selective by design. Activates when multiple environmental and market stress thresholds breach simultaneously. Historically observed during VIX ≥ 35, SPX declines ≥ 3% in 5 days or ≥ 6% in 10 days.

Precision
92%
Activations
13
Median lead
0 BD (same-day)

Institutional use: crisis playbook activation, pre-market risk committee awareness, scenario planning for extreme conditions.

2.2 SHOCK

Short-duration volatility expansion. Captures sharp, rapid VIX increases that typically resolve within a few trading days. Median episode duration: 3 business days.

Precision
85%
Activations
67
Median lead
0 BD (same-day)

Institutional use: tactical hedging awareness, portfolio rebalancing context, short-window risk committee notification.

2.3 ELEVATED

Advisory-tier elevated stress. Wider net, lower precision. Conditions warrant heightened vigilance but have not reached critical thresholds. Episodes average 5 trading days.

Precision
76%
Activations
49
Frequency
~1.4/yr

Institutional use: early awareness for risk committees, intermediate classification between routine and critical, scenario planning before potential escalation.

2.4 SHIFTING

Regime transition classification. The system detects conditions moving away from STABLE, typically with a multi-day lead before market dislocation. Most frequent non-STABLE risk shape.

Precision
93%
Activations
69
Median lead
2 BD

Institutional use: early warning for portfolio positioning, hedge activation before spreads widen, transition context from STABLE to stressed conditions.

2.5 STABLE

Baseline classification. Active when no stress or transition rules are firing. Represents routine market conditions: approximately 87% of all trading days in the backtest window.

STABLE is not a prediction of low risk. The classification can shift to ELEVATED or higher within a single session. STABLE means the system currently sees no active stress signal.

2.6 WHIPSAW

Sub-state of STABLE. Activates when VIX rises 25%+ over five sessions while the system holds STABLE. The system has independent environmental data indicating the scare is probably noise rather than a genuine regime change.

An Event Risk Brief opens automatically and tracks whether the system or the market proves correct. Historically, 117 of 123 such episodes resolved without the system being wrong (95% accuracy).

Part of
STABLE SHAPES group (counts as routine for hedge calculations and calm-share statistics)
Daily delivery
Reads "WHIPSAW" rather than plain STABLE in email and API

2.7 RESOLVED

Sub-state of STABLE. Activates when the system returns to STABLE after a real escalation episode (ELEVATED, SHOCK, or CRISIS) while VIX remains elevated above baseline (VIX ≥ 20). The system is saying the crisis has ended before conventional volatility signals confirm it.

Historically, 52 of 71 RESOLVED classifications were correct: the crisis was genuinely over and the system did not re-escalate (73% call-level precision). When wrong (19 episodes), the system re-escalated within a median of 10 trading days, catching the resumption. Median VIX at RESOLVED classification: 25.9.

Part of
STABLE SHAPES group (STABLE + WHIPSAW + RESOLVED = ~87% of trading days)
Daily delivery
Reads "RESOLVED" rather than plain STABLE in email and API
Detailed event criteria (what counts as a hit for each risk shape) and scoring horizons are documented in Validation & Methods, Section 2.

3. Event Risk Briefs

Event-driven briefings issued when market conditions and the daily classification form a defined situation. Not a daily report. Issued only when there is something to document.

Report TypeTriggerWhat It Tracks
DivergenceVIX spikes or SPX falls while the system holds STABLEWhether the system or the market proves correct
EscalationThe system reclassifies from STABLE to a stress risk shapeThe transition, its market context, and historical analogs
Stand-DownThe system returns to STABLE after a stress episodeThe full cycle from departure to resolution

Each report type opens when its trigger condition is met and closes when the situation resolves.

3.1 How Reports Work

  • Permanent numbering. Each report receives a numbered identifier (e.g. 2026-001) and a permalink in the archive.
  • Status tracking. Reports are either ACTIVE (developing) or CONCLUDED (resolved, outcome recorded).
  • Classification preserved. Every report records the system's classification at time of issue, so the track record is verifiable after the fact.
  • Accumulating record. The archive grows over time as a numbered track record of every situation the system classified and what followed.

4. How Classifications Work

Classifications are produced by a rules-based engine using versioned rule definitions. One classification per trading day, published pre-market (~07:30 ET).

Rule definition
Each risk shape has one rule that defines trigger conditions, episode windowing, and scoring horizon.
Winner classification
When multiple rules fire on the same day: CRISIS > SHOCK > ELEVATED > SHIFTING. Otherwise STABLE.
STABLE group
STABLE is the baseline. WHIPSAW and RESOLVED are sub-states within the same group (~87% of trading days).
Auditability
Metrics from versioned manifests (1990-2025). Rule IDs, versions, and validation files are traceable.
Full methodology, overfitting controls, data sources, and validation protocols at mindforge.tech/validation-and-methods.
Market Risk Classifier and related services are for research and informational purposes only. Not investment, financial, or trading advice. Past performance does not guarantee future results. Mindforge is not a registered investment adviser. See full terms at mindforge.tech/terms.
Documentation HubRequest Pilot

Research use only. Not investment advice. Past performance ≠ future results. Mindforge is not a registered investment adviser. Full terms · Methods

Risk Shapes Reference | Mindforge