On-chain analysis
On public blockchains anyone can see every transfer. Key figures can be built from that. But how much they tell you is very limited, because you rarely know who an address belongs to.
Learning objective: After this lesson you can judge what on chain metrics can show and where their limits lie.
Unlike with shares, on a public blockchain you can see every movement: how many addresses are active, how much is being moved and how long holdings have sat untouched.
Metrics from the blockchain
Key figures come from that. How many coins are on trading platforms right now? How long has nobody touched them? How much gain or loss would today's owners have if they sold now?
The real advantage
That sounds like a real advantage, and partly it is. With no share can you see how many units have just been sent to an exchange, perhaps to be sold.
The limit of attribution
The limit lies in attribution. An address isn't a person. One person can have thousands of addresses, and a single exchange address can belong to millions of customers. Anyone deriving how people behave from this is making assumptions they rarely disclose.
Common metrics include the number of active addresses, the volume of transfers, the age of holdings last moved and the ratio of market capitalization to so called realized value. For realized value, each coin is valued at the price at which it was last moved. That lets you estimate how much paper gain or loss all owners together currently have.
Matching addresses to owners
The biggest problem is assigning addresses to the right owners. There are rules of thumb for this, for example: if several addresses are used together in one transfer, they probably belong to the same person. Such rules are often wrong, and some users deliberately defeat them with privacy techniques. The number of active addresses therefore says neither how many people use a network nor how much economic activity there is.
A short data history
On top of that, the data only go back a few years. Since 2009 Bitcoin has gone through only a handful of big ups and downs. Thresholds for buy or sell signals are therefore set on very few examples. Claims like this metric was always this low at bottoms rest on four or five cases, and that's far too few for a reliable conclusion.
Summary
- All transfers are visible, but addresses aren't people.
- Active addresses say neither how many people use a network nor how much is happening.
- The data cover too few big cycles for reliable signals.
Did you get it?
What's the basic problem of on chain analysis?
Assigning addresses to the right owners. There are only rules of thumb for it, and they're often wrong.
Why isn't the number of active addresses a user count?
Because one person can have many addresses and one exchange address can belong to millions of customers.
Why are on chain signals statistically weak?
Because the data only cover a few big market cycles.
Check your understanding
Your task
Name two on-chain metrics from the lesson and write down for each which assumption about the owners of the addresses you need in order to read behavior from it. Explain in one sentence why an address is not automatically a person.
Sources and further reading
- Data provider Glassnode explains in its own research how the MVRV metric, which compares realized cost basis with current market value, is constructed and where its limits lie. View source ↗
Related
- Wallet, private key, and seed phraseStage 2
- Fundamental vs. technical analysisStage 4
- Reading a chartStage 4
- Matching questions for this stageQuestions