As Bitcoin pushes past $67,600 in October 2024, the question on every serious investor’s mind is where we stand in the current market cycle. While beginners rely on sentiment and price action alone, advanced crypto analysts combine multiple on-chain metrics to build comprehensive cycle phase detection systems. This tutorial walks you through constructing a multi-signal framework for identifying accumulation, markup, distribution, and markdown phases using publicly available blockchain data.
The Objective
This tutorial aims to help you build a systematic approach to cycle identification that goes beyond single-indicator analysis. By combining on-chain metrics, derivatives data, and macro indicators, you can create a weighted scoring system that assigns probabilities to each market cycle phase. The goal is not to predict exact tops and bottoms — an exercise in futility — but to identify high-probability zones where the risk-reward ratio favors specific portfolio actions.
The current market environment provides an ideal testing ground. Bitcoin’s post-halving price action, the surge in ETF inflows, and the growing institutional participation create a complex signal environment that rewards multi-dimensional analysis over simple trend-following strategies.
Prerequisites
Before building your cycle detection system, you need access to several data sources and a basic understanding of on-chain analysis concepts. Familiarity with Bitcoin’s UTXO model, an understanding of how exchange reserves affect price, and comfort with reading candlestick charts are essential foundations.
For data access, you will need accounts on at least two on-chain analytics platforms. Glassnode provides comprehensive on-chain metrics including MVRV, SOPR, and exchange flow data. CryptoQuant offers real-time exchange reserve tracking and miner revenue data. Both platforms offer free tiers with sufficient data for building an effective cycle detection system. Additionally, access to CoinMarketCap historical snapshots allows you to backtest your signals against actual price data across multiple cycles.
A spreadsheet application or basic Python knowledge will help you combine and weight the signals. The system described here can be implemented entirely in Google Sheets or Excel for those who prefer no-code solutions.
Step-by-Step Walkthrough
Step 1: Establish Your Base Indicators. Begin with the three foundational on-chain metrics for cycle analysis. The Market Value to Realized Value (MVRV) ratio compares Bitcoin’s current market capitalization to its realized capitalization — the value of all coins at the price when they last moved on-chain. Historically, MVRV readings below 1.0 indicate accumulation zones, while readings above 3.5 often precede distribution. As of October 2024, MVRV readings are trending upward from post-halving lows but remain well below distribution thresholds.
The Stock-to-Flow model, while controversial, provides a useful framework for understanding Bitcoin’s supply dynamics post-halving. The Puell Multiple, which compares daily miner revenue to the yearly moving average, helps identify when miner selling pressure may be approaching levels that historically precede cycle peaks.
Step 2: Add Exchange Flow Analysis. Track Bitcoin exchange reserves as a secondary signal. Declining exchange reserves suggest accumulation as investors move coins to cold storage, while rising reserves often precede selling pressure. Monitor the net flow of Bitcoin to and from major exchanges weekly. Sustained net outflows during periods of flat or declining price action represent one of the strongest accumulation signals available.
Combine exchange flow data with stablecoin supply metrics. Growing stablecoin market capitalization, particularly USDT and USDC, indicates dry powder waiting to be deployed. When stablecoin supply is expanding while Bitcoin exchange reserves are declining, the conditions for a markup phase are strengthening.
Step 3: Incorporate Derivatives Signals. The futures funding rate provides insight into market positioning. Persistently negative funding rates during price declines suggest overcrowded short positions and potential for short squeezes — conditions often seen near accumulation phase bottoms. Conversely, extreme positive funding rates during rapid price increases signal overleveraged long positions characteristic of distribution phases.
Open interest trends add another dimension. Rising open interest alongside rising prices suggests new capital entering the market (markup phase), while rising open interest with declining prices often indicates aggressive shorting (markdown phase). Declining open interest with flat prices suggests market participants are disengaging — typical of late accumulation.
Step 4: Build Your Scoring Matrix. Assign each indicator a score from minus 2 to plus 2, where negative scores indicate bearish signals and positive scores indicate bullish signals. Weight the indicators based on their historical reliability: on-chain metrics receive higher weight than derivatives data, which can be more volatile and prone to false signals.
A composite score above a certain threshold signals markup phase conditions, while scores below a negative threshold indicate markdown phase. Scores near zero suggest either accumulation or distribution, with the direction of exchange flows helping to distinguish between the two.
Step 5: Backtest Across Multiple Cycles. Apply your scoring system to historical data from the 2017, 2019-2020, and 2020-2021 cycles. Note how well the system identified phase transitions and where it produced false signals. Adjust the weighting factors based on which indicators performed best across multiple cycles rather than optimizing for a single historical period.
Troubleshooting
Common issues with multi-signal cycle detection systems include signal divergence — when different indicators point to different phases — and lag, where on-chain metrics confirm phase transitions only after they have occurred. When signals diverge, default to the more conservative interpretation and reduce position sizes until signals align. For lag issues, supplement on-chain metrics with real-time derivatives and order flow data to provide earlier signals.
Another frequent problem is overfitting the scoring weights to a particular historical cycle. Bitcoin’s market structure evolves with each cycle as institutional participation grows and new financial instruments (ETFs, regulated futures) change the dynamics. Regularly reassess your indicator weights and be willing to discard metrics that lose predictive power over time.
Mastering the Skill
True mastery of cycle phase detection comes from maintaining the system consistently across multiple complete market cycles. Keep a weekly journal of your indicator readings, composite scores, and the actual market outcomes. Over time, patterns will emerge that no single indicator can capture — the interplay between miner behavior, exchange flows, derivatives positioning, and macroeconomic conditions creates a rich tapestry that rewards patient, systematic analysis.
Share your findings with the on-chain analysis community on platforms like CryptoQuant’s commentary sections and Twitter’s technical analysis circles. Peer review and debate sharpen analytical frameworks more effectively than any individual refinement. As the crypto market continues to mature in 2024 and beyond, the analysts who combine rigorous on-chain methodology with adaptability to evolving market structures will be best positioned to navigate the cycles ahead.
Disclaimer: This article is for informational purposes only and does not constitute financial or investment advice. Always conduct your own research before making investment decisions.
weighted scoring across multiple on-chain metrics is the right approach. single indicator analysis keeps leading people astray.
agreed. i track 4 metrics and when 3 of 4 line up its usually right. the problem is patience, people cant wait for alignment
the 3-of-4 metric approach signal_stacker mentioned is basically what I use. problem is you can sit in cash for 8 months waiting for alignment while everything pumps 60%
OnChainOwl the weighted approach is solid but the weights themselves are subjective. curious how they handle regime shifts where old correlations break down
weighted scoring without backtesting across at least 2 cycles is just charting with extra confidence. agree with the skeptics here
weighted scoring without backtesting is astrology with math. the framework is sound but without historical validation its just educated guessing with extra steps
weighted scoring without a backtest is just a fancy opinion. Mika is right that whale behavior shifts between cycles, which makes historical validation even more critical
on-chain metrics shifted after spot ETFs launched. whale behavior pre-2024 was retail and early adopters. now its BlackRock moving size. the old models need recalibration
Raili J. ETFs changed the whale landscape entirely. pre-2024 on-chain metrics assumed retail dominated flows. thats no longer true
the post-halving framework here aligns with what Glassnode has been publishing. nice to see it in an accessible format.
wish the article included backtest results for the weighted scoring model. hard to trust a framework without historical validation.
quant_alpha exactly. the weighted model needs at least one full cycle backtest. without 2018 and 2021 data points the phase detection is just pattern matching on a single sample
fair point but backtesting on-chain models is tricky because the datasets themselves change over time. whale behavior in 2018 is different from 2024
exactly this. 2018 whale behavior was accumulation from ideologically motivated holders. 2024 whales are institutional funds with completely different patterns
Mika Laine is right that whale behavior shifts between cycles but thats exactly why you need a weighted multi-signal model. one indicator lying to you is noise, three lying at the same time is a signal
quant_alpha asking for backtests is valid but Mika Laine is right that on-chain datasets change retroactively. glassnode has restated MVRV calculations before
glassnode alignment is a good sanity check but their data has had gaps during high volume periods. would want to cross-reference with cryptoquant
metric_drift cross-referencing glassnode with cryptoquant is smart during high volume. seen data gaps in glassnode during peak volatility that would mislead any model
datavault_ the glassnode alignment is reassuring but also means the edge is shrinking. if everyone uses the same metrics they get priced in
BTC at 67k in october 2024 feels like centuries ago. we went to 108k and back to 70k since then. wonder how this framework would have scored the december peak
whale behavior in 2018 was accumulation from ideological holders. 2024 whales are institutional funds with completely different patterns. old models need recalibration
Elara K. exactly. the weighted model needs at least one full cycle backtest. without 2018 and 2021 data points the phase detection is just pattern matching on a single sample
cross-referencing glassnode with cryptoquant is smart during high volume. seen data gaps in glassnode during peak volatility that would mislead any model
Casper N. makes the best point in this thread. whale behavior shifting between cycles means any static model will eventually fail. need adaptive weights not fixed ones
cross-referencing glassnode with cryptoquant during high volatility is smart but both have restated historical data before. your backtest might be fitting noise
the 3-of-4 metric approach sounds great until you sit in cash for 8 months waiting for alignment while everything pumps 60%. model risk cuts both ways