PrimeScoreMarkets
PrimeScore / MarketsIn development

A clearer view
of volatility.

Markets move on events. Understand what changed, how unusual it is, and what it could mean for volatility.

Run the prototype Explore the platform

Event-driven intelligence for
options research and trading.

MARKET REPLAYHistorical snapshot
VIX · daily close37.32
05 FEB 2018Expansionrelative to prior history
Signed severity+0.9992
1,260 prior closes
Signed severity +0.9992 on a scale from minus one to plus one−1 Compression0Expansion +1
Scenario explorerExperimental
Scenario index46.64
Gap · index points+9.32
FRED / VIXCLS Recorded classifier output

Scenarios use an uncalibrated multiplier, not an independent volatility forecast.

Source & method

Severity and certainty were recorded from the repository's HTTP classifier. The slider calculates observed index × (1 + severity × certainty × k). Certainty is 1.0 for these snapshots; predictive performance is untested.

Inspect the sourced snapshot ↗

01 / The platform

Market events.
Structured context.

A price move tells you what happened. Understanding its significance takes history, a consistent scoring method, and a visible chain of reasoning.

The working prototypeObservation → classification → scenario
  1. 01

    Capture the observation.

    Sourced market history gives each event a reference point.

    VIXOVXHistorical replay
  2. 02

    Measure the deviation.

    The classifier returns signed severity, certainty and reasoning.

    Statistical scoringHTTP API
  3. 03

    Explore the implication.

    Vary an explicit assumption and inspect the volatility scenario.

    Visible equationLocal dashboard
/ A

Consistent classification.

Market and macroeconomic observations scored against their reference history.

/ B

Traceable reasoning.

Source provenance, certainty and the calculation behind each assessment.

/ C

Repeatable replay.

Return to the same observation, inspect its context, and reproduce the result.

02 / In development

Next: broader
market context.

The next stage connects broader event coverage to a continuous research workflow.

  1. 01

    Broader event coverage

    Cross-asset relationships and geopolitical events, alongside market and macro data.

    Planned
  2. 02

    Continuous market context

    Live ingestion, event history and corroboration across independent sources.

    Planned
  3. 03

    Evaluated volatility forecasts

    Calibrated models tested against held-out outcomes for dislocation research.

    Planned
03 / How it's engineered

Agentic engineering.
Explicit constraints.

AI agents work inside defined architecture and API boundaries. Runtime validation checks their changes; bounded retries keep failures visible.

Agentic AI systems operating under upfront constraints and runtime validation.

Explore the engineering harness
Control loop
  1. 01
    Define the boundariesArchitecture · interfaces · requirements
  2. 02
    Validate the changeTests · contract gates · fitness checks
  3. 03
    Resolve or escalateBounded retries · recorded decisions
Inspect the checks on GitHub
Build on the work

Let's compare notes.

Radu Pop, Founder & Engineer

Get in touch