Active addresses is the nearest thing available to a count of people genuinely using the network for something. What carries meaning is not the figure but its position against its own multi-year average.

That average divides a network that is growing from one that is shrinking, and this is the only line here that has survived contact with the history. Fixed levels have not, because the underlying figure climbs as adoption spreads.

What it actually measures

Above the average and still climbing is the healthiest arrangement this view produces. Both facts are needed: the level says where usage stands and the direction says where it is heading.

Above the average with momentum turning over is a different and more equivocal state. Usage remains historically strong while the impulse behind it has gone, which is where advances have tended to tire.

Recovering from a weak base is where sustained upturns have started. The level is still poor at that point, which is precisely why the direction matters more than the reading.

A drifting measure needs a drifting threshold

The obvious way to read any usage measure is to pick a level and call anything above it healthy. On a network that has grown by orders of magnitude, no such level survives contact with more than a few years of history.

Comparing against a long-run average solves that by letting the threshold move with the thing it is judging. What comes out is a statement about whether usage is strong for this network as it currently exists, which is the only version of the question that means anything.

New Address Growth Rate: a rising total can hide a fading market

Acceleration against a month earlier is the reading. Fresh participants arriving faster than they were describes genuine expansion, and arriving slower describes the opposite regardless of what the total is doing.

An accumulated count of participants travels in one direction only, which makes it comforting and largely useless. Enthusiasm can drain for months while the figure climbs steadily, because arithmetic leaves it no alternative.

Reading the pace instead removes the illusion completely. What emerges is whether the flow of arrivals is strengthening or weakening, and the total is structurally incapable of answering that question at all.

What it does not tell you

One address is not one person, and that is the great weakness here. Any individual may run dozens, while a single exchange address may stand behind thousands of customers, which makes the figure a stand-in and never a population.

It leads price sometimes and not dependably enough to act on. Usage can hold up straight through a fall or stay flat through an advance without either being a contradiction.

Automated activity is counted alongside genuine demand. Anything generating addresses at volume inflates the reading regardless of whether a person was involved.

How to read it

Healthy expansion. Usage sits above its long-run average and is still climbing, which is the strongest arrangement here.

Above trend, fading. Still historically strong, though the momentum behind it has turned over.

Constructive turn. Climbing off a depressed base, which is the ground durable upturns have tended to begin on.

Network contraction. Usage sits beneath its long-run average and is still falling.

Active Addresses is one of the views on the Mining & Network dashboard, along with New Address Growth Rate, NVT Ratio and Yardstick.

Common questions

Why the multi-year average as the threshold?

Because the underlying figure climbs as adoption spreads, which leaves any fixed level obsolete within a few years. A moving reference is what has historically told growth apart from decline.

Is one address one person?

No, and that is the great weakness here. Any individual may run dozens, while a single exchange address may stand behind thousands of customers. Follow the direction of travel and never treat the figure as a population.

Does it lead price?

Sometimes, and not dependably enough to act on. Usage can hold up through a fall or stay flat through an advance without either being a contradiction.

Does automated activity count?

It does, and nothing separates it out. Anything generating addresses at volume lifts the reading whether or not a person was involved.

Which matters more, level or direction?

Both, and neither alone. The level says where usage stands and the direction says where it is going, and the four states here come from combining them.

Why the pace instead of the count?

Because a network that keeps growing keeps producing addresses even once enthusiasm has drained. The pace settles whether participation is gathering or draining, which an accumulated figure never can.

Is the flat state uninformative?

Not at all. Onboarding neither gathering nor falling away is an equilibrium, and equilibria in this market have tended not to last.