About this tool

A regime indicator for discretionary traders. Tells you what kind of market you're in — not what to buy or sell.

What is this?

This dashboard runs a Hidden Markov Model — a statistical method used in speech recognition, bioinformatics, and finance — to classify the market into one of three regimes each trading day:

Each day, the model outputs the probability that the market is in each regime, given everything it has observed up to that day's close. Those three numbers always sum to 100%.

It's a regime indicator — a thermometer. Not a buy/sell signal. The model doesn't predict future prices; it tells you what kind of environment you're in right now so you can size positions and manage risk accordingly.

Why does it exist?

Discretionary traders often miss the moment a market regime changes. Vol gets quietly higher, correlations break, and the strategies that worked last quarter stop working — but it's hard to see in the moment. This tool is a second pair of eyes on the regime question: a quantitative, dispassionate read on what kind of market you're trading today.

It's built specifically for the four global indices that matter most to the author's portfolio context: Nifty 50 (India, primary), S&P 500 (USA, for global risk-on/off context), KOSPI (South Korea, an early-cycle Asian indicator), and Shanghai Composite (China, the swing factor for emerging markets).

How to use it

  1. Open the dashboard. Pick the index you care about from the dropdown.
  2. Read the three percentages. They sum to 100%. The largest one is the dominant regime.
  3. Layer it on your own analysis:
    • Bear > 70%  →  reduce position size, raise cash, avoid new entries, consider hedges.
    • Bull > 70%  →  momentum is friendly, can size up, trend-following strategies typically work.
    • Neutral > 70%  →  mean-reversion environment. No edge from this model; use other inputs.
    • No regime above ~60%  →  transition period. Watch closely; reduce conviction.
  4. Watch how it changes over time. Use the 90-day chart to see how confidently the model is in each regime. A drift from one regime to another, visible over 1–2 weeks, is more useful than a single day's reading.
  5. Check the table of historical regimes. See how long the current and past runs lasted, and what the index actually did during them.
What it is not: a trade signal. The dashboard never says "buy" or "sell". Treat the probabilities as one input among many in your discretionary process. The model has a 1–3 day lag at regime transitions — fundamental to HMMs, not a bug.

What's under the hood

For each index, a 3-state Gaussian Hidden Markov Model is trained on 16 years of daily data (2010 to today). The model learns from these features:

The model is retrained every Sunday on the full dataset including the just-finished week. Daily probabilities are recomputed Monday-Friday at 16:30 IST (after NSE close) using the latest trained model.

For technical detail, see the Methodology page.

About the author

AS

Aditya Sahasrabuddhe

Discretionary trader · builder of quantitative tooling

Aditya is an active trader managing a multi-account family portfolio in Indian equities. He builds analytical tools — including a daily AI-driven CIO brief and this regime indicator — to inform his own discretionary process rather than automate it. This dashboard is one piece of that workflow, shared publicly so other traders can use it (or argue with it) as they see fit.

LinkedIn

Important notice

This tool is provided for educational and informational purposes only. It is not investment advice. The author is not a SEBI Registered Research Analyst. Trading and investing in equities carries substantial risk of loss. You are solely responsible for your own decisions. Read the full disclaimer before using.