A free, real-time tool for detecting cryptocurrency market regimes using quantitative statistical models — built for traders, researchers, and anyone who wants to go beyond candlestick charts.
BTCMonitors.com applies Hidden Markov Models (HMM) — a class of probabilistic state-space models widely used in quantitative finance — to live cryptocurrency price data. The result is a real-time classification of the current market into one of four distinct regimes, updated every 30 seconds.
Unlike simple moving-average crossovers or RSI signals, an HMM captures the underlying statistical structure of market behavior: the tendency for markets to persist in a regime, the probability of transitioning between regimes, and the distribution of returns and volatility within each regime.
The site covers Bitcoin and the top 30 cryptocurrencies by market cap, each modeled with parameters calibrated to that asset's historical volatility profile.
The model uses Baum-Welch (EM) algorithm to estimate parameters from historical price data, and the Viterbi algorithm to decode the most probable current regime from recent returns. The Forward algorithm computes real-time regime probabilities. Learn more on the Learn page.
7-day hourly price history with each candle colored by the regime active at that timestamp — making regime transitions immediately visible.
Rolling return standard deviation chart showing how volatility evolves relative to the HMM's regime boundaries.
Histogram of recent hourly returns plotted against the Gaussian emission distributions for each regime state.
Time-series of the HMM Forward algorithm output — the probability of being in each regime at each point in time.
All price data is sourced from CoinGecko's public API and is subject to their terms. Data is for informational purposes only.
Hidden Markov Models for financial regime detection have a strong academic foundation. The use of regime-switching models in financial markets was formally established by James Hamilton's 1989 paper "A New Approach to the Economic Analysis of Nonstationary Time Series and the Business Cycle" (Econometrica), which demonstrated that macroeconomic time series are better described by regime-switching processes than by linear models. This framework was subsequently extended to financial markets and volatility modelling by Bulla & Bauer (2006) and others working on HMM applications to equity and currency markets.
BTCMonitors applies a 4-state Gaussian HMM to cryptocurrency markets specifically, building on this academic tradition with parameters tuned to Bitcoin's characteristic volatility profile. The choice of 4 states was validated using Bayesian Information Criterion (BIC) model selection — comparing 2-state through 6-state models and selecting the specification that best balances goodness-of-fit with model parsimony. The 4-state model outperforms simpler specifications by capturing the empirically distinct High Volatility Bear state (which a 3-state model conflates with the Low Volatility Bear) while remaining identifiable with available data.
The model runs entirely in the browser using client-side JavaScript — all HMM computation is transparent and reproducible. No data is sent to external servers for regime computation; the Forward Algorithm, Viterbi decoding, and Baum-Welch parameter estimates are all executed locally in real time using live price data from the CoinGecko public API.
Based on Hamilton (1989) regime-switching framework and Gaussian HMM theory. State count validated via BIC model selection across 2–6 state specifications.
Out-of-sample classification accuracy of 70–85% measured by regime persistence and return distribution alignment. Performs best during persistent Bull and High Vol Bear states.
All HMM computation runs in your browser. No proprietary black box — the Baum-Welch parameters, transition matrix, and Forward Algorithm output are all visible in the dashboard.
Current model uses 168-hour (7-day) lookback, 30-second refresh, per-coin volatility multipliers for 29 altcoins, and Viterbi + Forward Algorithm dual-output for both path and probability classification.
For a detailed walkthrough of the mathematical implementation — including the Baum-Welch EM steps, Forward Algorithm recursion, and Viterbi backtracking — see the Learn page or read the blog article How Bitcoin's Hidden Markov Model Detects Regime Changes.
BTCMonitors is designed to be useful across a range of experience levels and use cases:
The tool is free to use with no account required. Regime change alerts are also free — paid tiers exist for higher-frequency alerts and API access.
The BTCMonitors blog publishes analysis on regime detection methodology, Bitcoin market structure, and practical applications of quantitative signals in crypto trading:
The information on BTCMonitors.com is provided for educational and informational purposes only. It does not constitute financial advice, investment advice, trading advice, or any other type of professional advice. Cryptocurrency markets are highly volatile and speculative. Past regime patterns do not guarantee future results.
Always conduct your own research and consult a qualified financial advisor before making investment decisions. BTCMonitors.com is not responsible for any financial losses incurred from using this site.
Have a question, found a bug, or want to suggest a feature? We'd love to hear from you. Visit our Contact page.