BCSI · Historical Similarity
Bitcoin Cycle
Similarity Index
A multi-timeframe quantitative framework measuring how closely Bitcoin’s current four-bar return structure resembles homologous historical market cycles.
Indicator primer
What the BCSI measures
The Bitcoin Cycle Similarity Index asks a narrow question: how closely does the market structure visible now resemble comparable structures observed in earlier Bitcoin cycles? It evaluates that question independently across daily, weekly, monthly and yearly horizons, so a high similarity reading on one timeframe does not automatically imply a high reading on the others.
Each comparison uses completed return paths and separates two dimensions of resemblance: shape, which captures the geometry of the path, and amplitude, which captures the magnitude of the move. The resulting score is normalized on a 0–100 scale and is designed to make historical analogy measurable rather than anecdotal.
BCSI is descriptive, not deterministic. Historical windows that look similar can diverge immediately afterward, and the model does not treat similarity as a calibrated probability of a future return. Its purpose is to identify comparable historical structures, expose how strong that resemblance is, and keep the uncertainty around historical analogies visible.
QuantMetrics Lab publishes quantitative research and informational content only. Nothing on this page constitutes personalised investment advice or a recommendation to buy, sell or hold any asset. Model outputs are subject to uncertainty and should not be used as a standalone decision rule. Full disclaimer →
Pattern comparison & historical continuation
Current path versus historical references
The current four-bar path runs from −4 to 0 and stops at the latest completed bar. Historical reference paths use the same relative positions and continue from +1 to +4 to show what actually followed those matched historical windows.
Current structure
Multi-timeframe similarity
Current BCSI, Shape Similarity and Amplitude Similarity are shown together for all available timeframes. All three measures use the same 0–100 scale.
Official historical series
BCSI through time
Select a timeframe to inspect the history of the calculated BCSI and its two components. Hover or tap the chart to read individual observations.
Methodological framework
Four bars, two similarity dimensions
For each timeframe, the current four completed bars are compared with available homologous four-bar windows from previous cycles. References are re-ranked by recency at each observation and then combined using normalized recency weights.
Shape similarity
A cosine-based comparison measures how closely the direction and relative pattern of the four returns align.
Amplitude similarity
A magnitude-based comparison measures how similar the overall scale of movement is across the two return structures.
Dynamic references
Only valid homologous references are used, with available cycles weighted dynamically by recency.
Shape and amplitude similarity are combined using fixed model parameters within a normalized 0–100 scale. The public page describes the model logic and interpretation; implementation-level parameterisation is maintained within the underlying QuantMetrics Lab research and production specification.
| Timeframe | Comparison window | Anchoring | Current-data rule |
|---|---|---|---|
| Daily | 4 completed daily bars | Halving-cycle aligned | Current day excluded |
| Weekly | 4 completed 7-day bars | Halving-cycle aligned | Incomplete bar excluded |
| Monthly | 4 completed 30-day blocks | Halving-cycle aligned | Incomplete block excluded |
| Yearly | 4 completed calendar years | Calendar years | Current year excluded |
Data and maintenance
Operational specification
| Update frequency | Daily program execution |
| Market series | BTC-USD market data retrieved through yfinance |
| Current model | BCSI_MTF_v1.0 |
| Current software | Version 2.5.1 |
| BCSI window | Four completed bars per timeframe |
| Reference weighting | Dynamic recency ranking among available homologous references |
| Forward outcomes | Descriptive only; not used to determine the BCSI reading |
| Integrity status | passed |
Limitations
What the BCSI does not claim
No deterministic repetition
A similar historical window does not imply that the subsequent path will reproduce the earlier cycle.
Small cycle sample
Bitcoin has only a limited number of completed modern halving cycles, restricting the number of homologous references.
Timeframe dependence
Daily, weekly, monthly and yearly readings answer different questions and may diverge substantially at the same time.
Not a forecast
Forward historical outcomes summarize what followed prior references; they are explicitly excluded from the BCSI formula.
Research and access
Documentation and data access
BCSI conceptual and validation research
The Research Library records the conceptual foundation of BCSI, including its multi-timeframe similarity design, historical-reference logic, interpretation boundaries and the validation work associated with the indicator.
Open related research →Historical series and structured outputs
Extended historical series, versioned files and machine-readable outputs are available on request for research, institutional and licensing purposes.
Request data access →The Bitcoin Cycle Similarity Index is a quantitative research output, not personalised investment advice, financial advice, a recommendation, solicitation, price target or trading instruction. Historical similarity and subsequent historical outcomes do not imply that future Bitcoin market paths will repeat, and the BCSI should not be used as a standalone decision rule. Read the full Research & Investment Disclaimer.