TL;DR: On-chain analysis looks at public blockchain data, such as flows of crypto to and from trading platforms and the balances of large addresses ('whales'). Analysts use it as an extra clue for buying or selling pressure. The data is incomplete and can be misread; it is not a reliable predictor of price.
What can you expect?
- Time per week
- A few hours per week for looking at data and keeping a journal. Learning how the measures work first takes more time.
- Starting capital
- There is no fixed minimum. Use only money you can lose entirely. Crypto can fall fast and hard.
- Typical setbacks
- Signals that do not work out, mislabelled data, sudden price moves against your position and emotional decisions. You can lose your entire stake.
- Costs
- Trading costs and spread at your provider, network fees when moving crypto, and subscription fees of data providers. Compare provider costs.
- Not suitable if…
- Not suitable if you cannot tolerate large price swings, need money you cannot afford to lose or expect data to predict price reliably.
How it works
A blockchain is public. That lets you see how large amounts of bitcoin or other crypto move between addresses and trading platforms. Analysts call large holders 'whales' and look at patterns such as net inflows or outflows at exchanges. The idea is that this says something about selling or buying pressure. The problem: the data is incomplete, addresses cannot always be linked to an owner, and a transfer says nothing about what the holder will do. Use this as supporting information, not as a trading signal. Crypto is very risky and you can lose your entire stake.
Core principles
- 1. Use on-chain data as supporting information, never as the only reason to trade
- 2. Look at patterns over a longer period, not at a single transaction
- 3. Be aware that address labels (who owns it?) are estimates by data providers
- 4. Set your risk and stop in advance; crypto can move hard and fast
Step-by- step plan
- 1
Choose reliable data sources
Use one or more on-chain data providers and read how they label addresses. Data from different providers can differ.
- 2
Follow net flow over a longer period
Look at net inflow and outflow at trading platforms per week or month, not per hour. Watch for deviations from normal variation.
- 3
Compare with price and volume
Check whether the pattern coincides with price moves and volume. A signal without confirmation is weak.
- 4
Set risk and position size
Set your maximum loss and stop. Keep positions small and use no leverage.
- 5
Review after each trade
Record in a journal whether the on-chain signal added anything. That shows whether the approach is useful in your situation.
In detail
What is a 'whale' and what can you see?
In crypto, a holder of a large amount of coins is called a whale. There is no fixed threshold; for bitcoin, addresses with a thousand or more bitcoin are often cited.
On a public blockchain, transactions and address balances are visible. You see only addresses, not names. Data providers try to group addresses and link them to parties, for example an exchange. That link is an estimate and can be wrong.
A large transfer can therefore be a sale, but also an internal move, a custody change or a handover.
Flows to and from trading platforms
A common measure is net flow (netflow): what arrives at trading platforms minus what leaves. The reasoning is that crypto on an exchange can be sold faster than crypto in a personal wallet. Net inflow is therefore sometimes seen as possible selling pressure, net outflow as possible accumulation.
This is a hypothesis, not a rule. People also move crypto for other reasons: collateral, lending, switching providers or security. Exchanges themselves also sometimes move large amounts between their own wallets.
Look at patterns over weeks rather than one day, and compare several sources.
Combining with other data
On-chain data is often combined with other data, such as price action, volume and the funding rate and open interest of futures. A signal confirmed by several sources is usually more reliable than a standalone signal. But even then it stays uncertain.
Much data is only seen as 'clear' in hindsight. So test an approach on historical data and a demo account. Beware of stories about successful predictions: they rarely show the failed ones.
Limits and risks
The main limits: the data is incomplete, labels are estimates, and trading platforms sometimes hold client funds in pooled wallets. A move can therefore look larger or smaller than it is. A signal for bitcoin also does not automatically hold for other crypto.
Crypto-assets are highly volatile and are not covered by a deposit guarantee. The European MiCA regulation sets requirements for crypto-service providers, but does not protect against price losses. You can lose your entire stake. Use only money you can afford to lose.
- Basic understanding of blockchain and wallets
- Experience with crypto price charts
- Acceptance of high volatility
- On-chain data provider
- Blockchain explorer
- Charting platform
- Trading journal
- 01 Several independent sources show the same pattern over several days or weeks
- 02 The signal is confirmed by price, volume and other data (such as futures funding rates)
- 03 You know the limits of the data provider, such as how addresses are labelled
- 04 You have set your maximum loss and stop in advance and the position is small
- 01 The stop is hit: close without debate
- 02 The pattern in the data reverses or disappears
- 03 Price moves against your signal and does not confirm your assumption
- 04 You reach a period chosen in advance
Worked example
Hypothetical example (all quantities are assumed). You follow bitcoin's net flow at trading platforms over one week.
- Inflow to trading platforms: 12,000 BTC
- Outflow from trading platforms: 15,500 BTC
- Net flow = inflow - outflow = 12,000 - 15,500 = -3,500 BTC (net outflow)
A negative result is seen by some analysts as possible accumulation. That is a hypothesis.
Put the size in perspective. Suppose 19,000,000 BTC are in circulation: 3,500 / 19,000,000 = 0.018% of supply. Part of the outflow may also be an internal move.
What this means for you: the data gives at most an extra clue. Decisions on position size and stop remain the most important. Example: capital €5,000, maximum loss 1% = €50. If you buy at €60,000 per BTC with a stop at €57,000 (€3,000 risk per BTC), you may buy at most 50 / 3,000 = 0.0167 BTC (value about €1,000). Use the position size calculator for your own numbers.
This is a calculation example. It is not an expectation and not advice.
Hypothetical example, not advice.
Work it out yourself
Risks and risk management
- Incomplete or mislabelled data: a move between one party's own wallets looks like a large transfer
- Misinterpretation: a flow to an exchange does not always mean selling
- False signals and lag: the move may already be in the price
- Crypto is very volatile and there is no deposit guarantee; you can lose your entire stake
- Use small positions and set your maximum loss in advance
- Do not use leverage on uncertain signals
- Hold crypto with a provider authorised under MiCA or securely in self-custody, and mind hack risk
- Keep a journal to test whether the signal helps in your situation
Learn the basics
Key takeaways
- A lot is visible on a public blockchain, but who owns an address and what they will do remains an estimate.
- Net inflow or outflow at trading platforms is a common measure, but gives no certainty about price.
- Combine signals and limit position size. A stop is mandatory.
- Crypto has no deposit guarantee and can fall sharply and fast.
Frequently asked questions
Where do I find data on whale movements? +
There are on-chain data providers (with basic and paid plans), blockchain explorers and alerts for large transactions. Quality and labelling methods differ. Compare sources and read how a measure is calculated.
How do I tell accumulation from an ordinary transfer? +
You cannot with certainty. Look at patterns over several weeks and at several sources. A single large outflow can also be an internal move by a trading platform. Treat every signal as a hypothesis.
Does this also work for Ethereum or other crypto? +
Comparable data exists for other crypto, but is often less complete or less reliably labelled. There is no guarantee that a pattern in bitcoin holds elsewhere. Extra caution is in order. Crypto-assets are not covered by a deposit guarantee.
Origin
General technique