Regime · S&P 500 Volatility Regime

P(Low Vol)–
P(Mid Vol)–
P(High Vol)–
Days in Regime–
VIX–
Realized Vol–
As of–

S&P 500 by Regime

Low VolMid VolHigh Vol

Regime Map

VIX vs realized volatility, each dot is one session

Regime Probabilities

Model probability of each regime by day

Regime Statistics

Since 2007, out of sample

By regime

Next-day transition probability

S&P 500 volatility regime today

This page classifies every trading day since 2007 into one of three volatility regimes for the S&P 500, Low Vol, Mid Vol and High Vol, using a proprietary unsupervised machine-learning model built on historical S&P 500 volatility surface data. It shows today's regime and its probability, how the index behaved in each regime, and how often one regime turns into another.

How to read this page

Methodology

The regime model is a proprietary unsupervised machine-learning model developed by QuantDaddy. In broad terms it works in four steps:

  1. Inputs. The model uses indicators derived from historical S&P 500 implied volatility surface data, together with index price history, to describe the volatility state of each session.
  2. Learning the regimes. The model learns three recurring market states from history without being told in advance what calm or stressed markets look like, and orders them from lowest to highest volatility as Low, Mid and High Vol.
  3. Walk-forward, no look-ahead. The model is trained on earlier history and then rolled forward one session at a time from 2007. Each day is classified only with information available at that time, and the model is periodically re-trained as new data arrives, so the history shown is out of sample.
  4. Daily output. For every session the model produces a probability for each regime; the regime shown is the most likely one. When the latest session is newer than the last model run, a preliminary estimate is shown and replaced by the final value after the daily update.

Model specification and parameters are not published.

Volatility regime FAQ

What is a volatility regime?

A volatility regime is a persistent market state with its own typical level of volatility and behavior of returns. Equity markets tend to alternate between long calm periods with low volatility and steady gains, transition periods, and shorter stress periods with high volatility and large daily swings. Identifying the current regime helps put GEX, the VIX term structure and option prices in context.

How is the regime determined?

A proprietary machine-learning model learns three typical market states from indicators built on historical S&P 500 volatility surface data. For each session it estimates the probability of belonging to each state, and the regime shown is the one with the highest probability.

Is the history re-written with hindsight?

No. From 2007 onward each day is classified by a model trained only on data available before that day, and the model is re-trained periodically as new data arrives. Past labels shown here are the labels the model would have produced at the time.

Does the regime predict where the market is going?

No. The regime describes the current volatility state, not the direction of the next move. Regimes are persistent, so today's regime is a reasonable guide to tomorrow's, but transitions can be abrupt, especially into High Vol. Use it as context, not as a trading signal.

Why can probabilities change sharply?

A single large move changes both realized and implied volatility, which can push a session across the boundary between two regimes. Periodic re-training can also shift those boundaries slightly.

When is it updated?

Once a day after the US close, with the latest market data. During the session, or before the daily update has run, the page shows a preliminary estimate for the latest session.

Learn more

Market data from Cboe. Model output is statistical and can be wrong. Nothing here is investment advice.