| Indicator | Value | Signal |
|---|
Bitcoin Market Regime Detection with Hidden Markov Models
BTCMonitors applies a 4-state Gaussian Hidden Markov Model (HMM) to Bitcoin's hourly price data to identify the current market regime in real time. Unlike price-based indicators that react to individual candles, the HMM models the full statistical distribution of recent returns — detecting when the market has genuinely shifted into a different behavioural state, not just when price has crossed a threshold.
The model is trained on years of Bitcoin price history using the Baum-Welch algorithm, an iterative expectation-maximisation method that learns the characteristic return distribution for each of the four market states. Real-time classification uses the Forward Algorithm to produce continuous probability estimates across all four states and the Viterbi Algorithm to decode the single most probable state sequence. The dashboard updates every 30 seconds, classifying 168 hourly observations (7 days) on each refresh.
The result is a quantitative confidence score — not a binary buy/sell signal — that tells you both which regime Bitcoin is most likely in and how certain the model is. A 91% confidence Low Volatility Bull reading means the statistical characteristics of recent returns are overwhelmingly consistent with a bull regime. A 58% reading means genuine uncertainty between two states — itself an important signal to reduce leverage and wait.
The Four Bitcoin Market Regimes
Consistent positive hourly returns with low variance. Bitcoin is trending steadily upward. The model's most favourable regime for long exposure — positive expected returns with reduced volatility risk.
Negative mean returns with moderate volatility. The "slow bleed" — consistent selling pressure without dramatic crashes. Often the longest-duration bear regime and the most corrosive for leveraged positions through funding costs.
Large negative returns with very high variance. Crash and capitulation conditions. Extreme intraday swings in both directions, liquidation cascades, and widening spreads. Typically the shortest regime but the most damaging without early detection.
Near-zero mean with moderate variance. The market is genuinely between regimes — the Forward Algorithm assigns meaningful probability to multiple states simultaneously. High uncertainty periods that often precede sharp directional moves.
How to Use This Dashboard
The Current Regime panel shows the active state and its confidence percentage. A confidence above 75% indicates a statistically dominant regime — the primary threshold for acting on a regime signal. The probability bars show the full distribution across all four states, which is more informative than the leading label alone during Transition or low-confidence periods.
The Regime Signals table aggregates indicators including realized volatility, the volatility spread between regimes, and forward probability trends to produce a composite signal summary. The Viterbi Decoded Path shows the most probable state sequence across the last 60 periods, providing historical context for the current classification. For a deeper guide on interpreting these signals, read Using HMM Regime Signals in Your Strategy →
Frequently Asked Questions
Bitcoin regime detection identifies hidden statistical market states — such as Low Volatility Bull, High Volatility Bear, or Transition — using quantitative models applied to price return data. Unlike technical indicators, regime detection does not require a price to cross a threshold; it classifies the statistical character of recent market behaviour as a whole.
An HMM is a probabilistic model that infers hidden states from observable data sequences. Applied to Bitcoin, the hidden states are the market regimes and the observable data is the sequence of hourly log-returns. The model learns the characteristic return distribution for each regime and uses this to compute a probability that the market is currently in each state, updated continuously as new price data arrives.
The regime classification updates every 30 seconds. On each update, the latest hourly return data is fetched from CoinGecko, the 168-observation (7-day) lookback window is recalculated, and the Forward Algorithm recomputes the state probability distribution. The displayed regime reflects the state with the highest probability at that moment.
BTCMonitors tracks Bitcoin on the main dashboard and provides individual HMM regime dashboards for 29 additional cryptocurrencies including Ethereum, Solana, XRP, BNB, DOGE, ADA, AVAX, LINK, DOT, and more. Each altcoin uses the same 4-state HMM framework with a per-coin volatility multiplier calibrated to that asset's historical return distribution. Access all coin dashboards via the Markets menu in the navigation.
The BTCMonitors Alerts feature sends an email notification within minutes of a confirmed regime change. You can configure alerts for any regime transition, or specifically for High Volatility Bear detection — useful for receiving early warning of crash conditions. Alerts are free and require no account creation.