S&P 500 Risk Regime

Your agent reads today's regime flag, then pulls the history of flags to describe how volatility, rather than direction, has behaved after similar days.

One flag for the state of the market.

A risk regime is a coarse description of market conditions. This dataset carries one binary label per trading day: 1 when conditions are risk-off, meaning elevated volatility or stress, and 0 when they are risk-on, meaning lower volatility and constructive equity conditions.

The label is derived from forward-looking implied volatility metrics rather than from what the market has already done, which is what makes it usable as an input on the day instead of a description of last week.

Size exposure by regime instead of by mood.

The honest version of this signal is a volatility forecast. Across 3,833 trading days, risk-off days realized 24.7% volatility against 12.8% on risk-on days, and forward 10-day volatility ran 19.0% against 12.3% with a t statistic of 15.0. Returns in the two states were 13.0% and 12.7%. The flag tells you how rough the ride will be, not which way it goes.

That makes it a sizing and volatility-timing input. Cut gross exposure when the flag is 1, buy convexity when it flips, stay out of short premium in the state where realized volatility doubles, and split every backtest by regime so a strategy that only works in calm conditions cannot hide inside a blended average.

• Scale position size with a regime multiplier instead of running a fixed risk budget.
• Segment an existing backtest on date and compare drawdown, not just return, across the two states.
• Keep short-volatility strategies out of risk-off days, where realized volatility roughly doubles.
• Treat the flip from 0 to 1 as a dated event and study what follows in your own book.
• Join the flag to any other dated dataset to ask whether an effect is regime dependent.

Read from implied volatility, fixed once published.

The classification is recomputed daily at 10:10 in New York from forward-looking implied volatility metrics, then written as a fixed daily label. Historical values are never retroactively altered, so a study over the history reads the same flag that was live that morning.

The published market-effect study joins 3,833 trading days back to 2011 against SPY, which is the sample every number on this page comes from. Pro includes full history and the latest session.

Two fields, plus the filters that pull the history.

The headline fields below are a subset. Every field, with exact types and semantics, is documented in the API reference.

One call returns years of daily regime labels.

One key works across the REST API, the hosted MCP server, and this dashboard. Every response is JSON with a { count, data } envelope.

import requests

url = "https://api.alphanume.com/v1/sp500-risk-regime"
params = {
    "date_gte": "2026-01-01",
    "date_lte": "2026-08-25",
    "api_key": "alp_abc123"
}

r = requests.get(url, params=params)
print(r.json())

curl "https://api.alphanume.com/v1/sp500-risk-regime?date_gte=2026-01-01&date_lte=2026-08-25&api_key=alp_abc123"

{
  "count": 2,
  "data": [
    { "date": "2026-08-25", "risk_regime": 1 },
    { "date": "2026-08-24", "risk_regime": 0 }
  ]
}

What this data does not claim.

• The flag is not directional, and the study says so in public: 13.0% against 12.7% annualized returns across the two states, with the flat chart published rather than dropped.
• It is binary. A 1 does not say how deep the stress runs, so a day that just crossed the line looks identical to a crisis day.
• It classifies current conditions. It is not a crash forecast, and most risk-off days are ordinary volatile days that resolve without incident.
• The row carries the label and nothing underneath it, so you cannot see how close a day came to flipping.
• The study is a market-level test against SPY. It says nothing about how the regime interacts with your specific strategy, which is a join you have to run yourself.