CONVEX
Glossary/Crypto & Digital Assets/On-Chain Metrics
Crypto & Digital Assets
8 min readUpdated Apr 12, 2026

On-Chain Metrics

ByConvex Research Desk·Edited byBen Bleier·
on-chain analysisblockchain analyticson-chain dataon-chain indicatorsblockchain dataGlassnode metrics

Data derived directly from the Bitcoin or Ethereum blockchain, including wallet flows, exchange balances, long-term vs short-term holder behaviour, and miner activity, offering a transparent view of supply and demand dynamics unavailable in traditional markets.

Current Macro RegimeSTAGFLATIONDEEPENING

The macro regime is unambiguously STAGFLATION DEEPENING. The hot CPI print (pending event, 24h ago) is not a surprise — it is a CONFIRMATION of the pipeline signals that have been building for weeks: PPI accelerating faster than CPI, Cleveland nowcast at 5.28%, breakevens rising +10bp 1M across the …

Analysis from May 14, 2026

What Are On-Chain Metrics?

Because blockchains are public ledgers, every transaction, every wallet balance, every transfer, every fee paid, is permanently recorded and queryable by anyone. On-chain analysis uses this unprecedented data transparency to understand market structure, investor behaviour, and supply/demand dynamics from first principles.

This is impossible in traditional finance. You cannot see how much gold is moving between institutional vaults in real time. You cannot observe the aggregate cost basis of all S&P 500 shareholders. You cannot track whether long-term holders of Treasury bonds are accumulating or distributing. In Bitcoin, you can see all of this, and it gives on-chain analysts an information edge that has no parallel in traditional markets.

The field was pioneered by pseudonymous analysts on Twitter (notably "Willy Woo" and "PlanB" in 2017-2019) and has since been institutionalised through analytics platforms like Glassnode, CryptoQuant, Santiment, and Coin Metrics. On-chain data now informs the investment decisions of hedge funds, family offices, and Bitcoin ETF managers.

The Core On-Chain Metric Categories

1. Exchange Flows, The Supply/Demand Pulse

Exchange flows measure coins moving to and from exchange deposit addresses. They are the single most immediately actionable on-chain signal.

Metric What It Measures Bullish Signal Bearish Signal
Exchange inflows BTC moving to exchanges Low/declining Spike (>10K BTC/day)
Exchange outflows BTC leaving exchanges Sustained high outflows Declining outflows
Exchange balance Total BTC on exchanges Declining (supply squeeze) Rising (sell-side builds)
Net flow Inflows minus outflows Negative (outflows > inflows) Positive (inflows > outflows)

The structural story: Bitcoin on exchanges declined from approximately 3.1 million BTC in March 2020 to under 2.3 million BTC by 2024, about 800,000 BTC ($40-60 billion) removed from exchange-accessible supply. This is the longest sustained exchange withdrawal trend in Bitcoin history and represents a structural supply squeeze that underpins higher prices.

Notable exchange flow events:

  • May 2021 crash ($58K → $30K): Preceded by 3 days of elevated inflows totaling ~40,000 BTC
  • FTX collapse (November 2022): Triggered the largest single-week exchange outflow in history as users withdrew to self-custody (~200,000 BTC net)
  • ETF accumulation (2024): BTC moved to Coinbase Prime and other custody solutions for ETF backing, technically exchange outflows that represent institutional buying

2. Holder Behaviour, Who Is Buying and Selling?

Glassnode pioneered the concept of Long-Term Holders (LTH), wallets that haven't moved BTC in 155+ days, versus Short-Term Holders (STH), coins acquired within 155 days. This classification reveals the market cycle with remarkable clarity.

LTH/STH supply cycle:

Cycle Phase LTH Supply Trend STH Supply Trend Price Action
Accumulation (bear bottom) Rising sharply Shrinking Flat/bottoming
Early bull High plateau Slowly rising Beginning to rally
Late bull / distribution Declining (selling to new entrants) Rising rapidly Parabolic rally
Capitulation Begins rising again Collapsing Crash/bottom forming

In the 2020-2022 cycle: LTH supply peaked at ~69% of circulating supply in early 2021, declined to ~58% by November 2021 (peak distribution), then rebuilt to ~76% by late 2023 (deepest accumulation on record).

STH Realized Price: The aggregate cost basis of short-term holders. When BTC trades below STH realized price, it means recent buyers are underwater and under maximum psychological pressure to sell, this has marked every major capitulation bottom since 2018. When BTC reclaims STH realized price from below, it confirms a new uptrend.

3. Profitability Metrics, The Psychology of Profit and Loss

Metric Formula Cycle Top Signal Cycle Bottom Signal
MVRV Ratio Market Cap ÷ Realized Cap >3.5 (avg holder up 250%+) <1.0 (avg holder at a loss)
MVRV Z-Score (MV - RV) ÷ StdDev(MV) >7 <0
SOPR Spent output value ÷ creation value Persistently >1.05 (profit-taking) Persistently <1.0 (capitulation)
NUPL (Market Cap - Realized Cap) ÷ Market Cap >0.75 (euphoria) <0 (capitulation)
Realized P/L Ratio Daily realized profit ÷ realized loss >10 for weeks (euphoria) <0.1 for weeks (despair)

MVRV in practice: MVRV peaked at 4.7 in December 2017, 3.96 in April 2021, and bottomed at 0.76 in November 2022 (post-FTX collapse). The convergence of MVRV to lower cycle peaks over time reflects Bitcoin's maturing investor base and declining volatility.

SOPR mechanics: When SOPR drops below 1.0 and stays there, it means coins are being spent at a loss, holders are capitulating. When SOPR recovers back above 1.0 after sustained sub-1.0 readings, it signals the capitulation is over and a new bull phase is beginning. SOPR briefly touched 0.93 during the FTX collapse, deep capitulation, before recovering in early 2023.

4. Miner Metrics, The Supply Side

Bitcoin miners produce all new supply (6.25 BTC per block through April 2024, then 3.125 BTC post-halving). Their behavior directly impacts sell-side pressure.

Metric What It Tracks Why It Matters
Hash rate Total mining computational power Network security; miner commitment
Miner revenue Block rewards + transaction fees Miner profitability determines selling pressure
Miner outflows BTC moving from miner wallets to exchanges Direct sell-side pressure
Hash ribbons 30-day vs 60-day hash rate moving averages Miner capitulation signal when inverted
Puell Multiple Daily miner revenue ÷ 365-day average revenue >4 = overbought; <0.5 = oversold

Hash ribbon capitulation: When the 30-day hash rate moving average crosses below the 60-day average, miners are shutting off machines, typically because BTC price has fallen below their cost of production. Historically, hash ribbon capitulation events have been among the best buying signals in Bitcoin: they occurred in January 2019 ($3,500 BTC), March 2020 ($5,000), June 2022 ($20,000), and December 2022 ($17,000). In every case, BTC was significantly higher 12 months later.

Post-halving miner stress: After each halving, miner revenue is instantly cut in half. Less efficient miners (those with higher electricity costs or older equipment) must sell BTC reserves to fund operations or shut down entirely. This creates a 3-6 month period of elevated miner selling that can suppress price in the near term but ultimately reduces future sell-side pressure as weak miners exit.

5. Network Activity, Demand Side

Metric Measures Interpretation
Active addresses Unique addresses transacting daily User adoption proxy
New addresses First-time addresses New user onboarding rate
Transaction count Daily on-chain transactions Network usage
Transfer volume USD value of on-chain transfers Economic activity
NVT Ratio Market cap ÷ daily transfer volume "P/E ratio for Bitcoin"

NVT Ratio: Willy Woo's Network Value to Transaction ratio is conceptually similar to a P/E ratio, it compares Bitcoin's valuation to the economic activity flowing through its network. When NVT is elevated (>95th percentile), BTC may be overvalued relative to its network usage. When NVT is low, the network is processing significant value relative to its market cap, suggesting undervaluation.

Building an On-Chain Dashboard

The "Must Watch" Short List

For traders who don't want to monitor dozens of metrics, these 5 provide the highest signal-to-noise:

  1. Exchange net flow (7-day average): Are coins flowing to or from exchanges?
  2. MVRV Z-Score: Are we at cycle extremes?
  3. LTH supply change (30-day): Are long-term holders accumulating or distributing?
  4. STH realized price: Are recent buyers in profit or at a loss?
  5. Funding rate (covered separately but complementary): What is derivatives leverage sentiment?

Data Sources

Platform Strengths Cost Best For
Glassnode Deepest metric library, institutional-grade Free tier limited; Pro $40/mo; Advanced $840/mo Serious analysts
CryptoQuant Exchange flow data, miner metrics Free tier decent; Pro $29/mo Exchange flow focus
Santiment Social + on-chain combined Free tier; Pro $49/mo Sentiment overlay
Coin Metrics Academic rigor, API-first Network Data Pro varies Quant/data teams
Blockchain.com Free basics, long history Free Quick checks

Limitations and Pitfalls

On-chain analysis is powerful but not infallible:

  1. Entity identification is imperfect: Analytics firms use heuristics to cluster addresses belonging to the same entity, but these can misfire. A single exchange moving funds between internal wallets can create false "whale movement" alerts.

  2. Wrapped and bridged BTC is invisible: BTC bridged to Ethereum as WBTC (~150,000 BTC) or locked in Lightning Network channels doesn't appear in standard on-chain metrics.

  3. ETFs create new blind spots: Bitcoin ETFs hold ~1 million BTC in institutional custody that moves differently than organic on-chain activity. Coinbase Prime movements serving ETF redemptions/creations are technically "exchange flows" but represent a fundamentally different dynamic than retail buying/selling.

  4. Historical patterns can break: Metrics calibrated against 2013, 2017, and 2021 cycles may behave differently as market structure evolves. The 2024 cycle's ETF-driven demand has no historical precedent.

  5. Survivorship bias: The metrics that "called" every cycle top also produced intermediate signals that are less discussed. MVRV reached concerning levels in mid-2019 at $13K, far from the actual cycle top.

What to Watch

  1. Exchange balance trend, the multi-year decline in BTC on exchanges is the most powerful structural bullish signal. If this trend reverses (exchange balance starts rising persistently), it would be a major bearish development.
  2. LTH/STH crossover, when LTH supply starts declining after an extended accumulation phase, distribution has begun and the clock is ticking on the cycle.
  3. MVRV extremes, use Z-Score rather than raw MVRV for cross-cycle comparisons. Values approaching 7+ warrant defensive positioning regardless of narrative.
  4. Hash ribbon capitulation, the single best "buy the blood" signal in Bitcoin's history, but requires patience (capitulation can last weeks to months).
  5. Stablecoin supply on exchanges, growing stablecoin reserves on exchanges represent dry powder waiting to deploy into BTC and alts.

Frequently Asked Questions

What are the most important on-chain metrics for predicting Bitcoin price?
The highest-signal on-chain metrics for price prediction, ranked by historical accuracy: (1) MVRV Z-Score — when Market Value exceeds Realized Value by 7+ standard deviations (Z-Score > 7), BTC has been within 1-3 months of a cycle top in every previous cycle (2013, 2017, 2021). When MVRV Z-Score drops below 0, BTC has been near cycle bottoms. (2) Exchange balance trend — BTC on exchanges declined from 3.1 million in early 2020 to under 2.3 million by 2024. This structural supply drain is the longest sustained withdrawal trend in Bitcoin history and supports higher prices by reducing available sell-side liquidity. (3) Long-term holder supply — when LTH supply (coins unmoved 155+ days) starts declining after a period of accumulation, it signals distribution into strength and typically precedes cycle tops by 3-6 months. (4) Realized profit/loss ratio — when daily realized profit exceeds realized loss by 10:1 or more for sustained periods, it marks euphoric distribution. Conversely, when realized losses dominate for months, it signals capitulation. (5) SOPR (Spent Output Profit Ratio) — values persistently below 1.0 confirm capitulation; the recovery of SOPR back above 1.0 after sustained sub-1.0 readings has preceded every major rally.
How do exchange flows signal buying or selling pressure?
Exchange flows are the on-chain equivalent of watching institutional order flow: (1) Exchange inflows (coins moving to exchange deposit addresses) represent potential selling pressure — the only reason to move BTC to an exchange is to sell, trade, or use as collateral. Large inflows (>10,000 BTC in a single day) are bearish signals, especially when concentrated in a few wallets. The May 2021 crash from $58K to $30K was preceded by 3 consecutive days of elevated exchange inflows totaling ~40,000 BTC. (2) Exchange outflows (coins leaving exchanges to private wallets) represent buying pressure that has been completed — someone bought BTC and moved it to cold storage for long-term holding. Sustained outflows exceeding inflows is structurally bullish. During the 2023-2024 accumulation phase, net exchange outflows averaged 2,000-5,000 BTC per week. (3) Exchange balance (total BTC held on all exchange addresses) is the aggregate picture. The exchange balance dropped from ~3.1 million BTC in March 2020 to ~2.3 million in 2024 — approximately 800,000 BTC ($40-60 billion) removed from exchange-accessible supply. This represents the largest sustained supply squeeze in Bitcoin history. Caveats: exchange wallet identification is imperfect (new wallets take time to label), and internal wallet movements can create false signals.
What is the MVRV ratio and how do I use it?
The Market Value to Realized Value (MVRV) ratio compares Bitcoin's current market capitalization to its "realized capitalization" — the sum of all BTC valued at the price when each coin last moved on-chain. Realized cap represents the aggregate cost basis of all Bitcoin holders. MVRV = Market Cap / Realized Cap. When MVRV > 1, the average holder is in profit; when MVRV < 1, the average holder is underwater. Historical cycle signals: MVRV peaked at 4.7 in December 2017, 3.96 in April 2021, and bottomed at 0.76 in November 2022 (after FTX collapse). The MVRV Z-Score normalizes this ratio against its historical volatility, making signals comparable across cycles. Z-Score > 7 has marked every cycle top. Z-Score < 0 has marked every cycle bottom. The metric works because it captures the aggregate psychology of the market: at high MVRV, holders have large unrealized gains and are incentivized to sell; at low MVRV, holders are at a loss and the marginal seller has been exhausted. Key limitation: MVRV treats lost coins (estimated at 3-4 million BTC) as "coins that moved at very low prices," which biases realized cap downward and MVRV upward. Despite this, the directional signal remains extremely reliable.
How do long-term and short-term holder dynamics signal cycle tops and bottoms?
Glassnode classifies Bitcoin into Long-Term Holder (LTH) supply — coins unmoved for 155+ days — and Short-Term Holder (STH) supply — coins moved within 155 days. The interplay between these cohorts maps the entire market cycle: (1) Accumulation phase (bear market bottom): LTH supply steadily increases as long-term buyers accumulate at low prices. STH supply shrinks as speculators exit. LTH supply peaked at ~76% of circulating supply in late 2023. (2) Early bull market: LTH supply remains high but flattens. New buyers increase STH supply. Prices begin rising. (3) Distribution phase (late bull market): LTH supply starts declining — long-term holders are selling to new entrants at higher prices. The rate of LTH supply decrease accelerates. When LTH supply drops by more than 5% of circulating supply over 3-6 months, the cycle top is near. In 2021, LTH supply declined from 69% to 58% between April and November — a massive distribution. (4) Capitulation: STH supply collapses as speculators sell at a loss. LTH supply begins rising again as the remaining holders refuse to sell and new "bottom buyers" accumulate. The STH cost basis relative to spot price is also critical: when BTC trades below the STH realized price (the aggregate cost basis of short-term holders), it signals capitulation and has marked every major bottom since 2018.
What are the limitations and pitfalls of on-chain analysis?
On-chain analysis is powerful but has important blind spots: (1) Entity identification is imperfect — analytics firms like Glassnode use heuristics to cluster addresses belonging to the same entity, but these can be wrong. A single exchange moving funds between internal wallets can create false "whale movements." Chainalysis and Glassnode have different entity labels, leading to different readings. (2) Wrapped and bridged BTC is invisible — BTC bridged to Ethereum as WBTC (~150,000 BTC) or used in Lightning Network channels doesn't appear in standard on-chain metrics. (3) OTC desks and ETFs create blind spots — large institutional trades via Genesis, Cumberland, or other OTC desks may not touch public exchanges. Bitcoin ETFs hold ~1 million BTC in custody that moves differently than organic on-chain activity. (4) Historical pattern breaks — on-chain metrics calibrated against 2013, 2017, and 2021 cycles may not work identically as market structure evolves (ETFs, institutional custody, derivatives dominance). (5) Survivorship bias — metrics that "called" every cycle top also produced false signals that are less discussed. MVRV reached high levels in mid-2019 that would have suggested a top, yet BTC was only at $13K. (6) Manipulation — sophisticated actors can create false on-chain signals by moving coins between self-controlled wallets to simulate accumulation or distribution patterns.

On-Chain Metrics is one of the signals monitored daily in the AI-driven macro analysis on Convex Trading. The platform synthesises data across monetary policy, credit, sentiment, and on-chain metrics to generate actionable trade recommendations. Create a free account to build your own signal layer and see how On-Chain Metrics is influencing current positions.

ShareXRedditLinkedInHN

Macro briefings in your inbox

Daily analysis that explains which glossary signals are firing and why.