What the number costs
Generating an onion address is close to free. A machine produces valid keys as fast as it can hash, and any of them can be published. So the cost of twelve is not in making them. It is in running them, because each address that answers has to be configured on a service that is listening, and each one is a separate descriptor being published to the network.
The other cost is on the reader. Three addresses can be held in mind well enough to notice when a fourth appears from somewhere unexpected. Twelve cannot. Nobody carries twelve 56 character strings in their head, which means every one of them has to be checked against a source every time, and the source becomes the thing that matters rather than the addresses.
What a bigger set buys
The obvious answer is redundancy. If addresses become unreachable for whatever reason, having more of them published means fewer people are stranded. That is the case usually made and it is not wrong.
The less obvious answer is that a large set is harder to displace. If a market publishes one address, an attacker who gets a lookalike in front of you is competing against one string. Against twelve, the lookalike has to displace the whole list, which is a bigger job than swapping one entry. A list is a heavier object than an address.
What it costs in the other direction
A long list dilutes attention. When there are three strings a reader might genuinely compare all three. With twelve, almost nobody checks the list they were given against the list they had. The natural behaviour is to grab whichever one is nearest and go, and that behaviour makes an extra entry easy to smuggle in.
A long list also makes ranking irresistible, and ranking is where lists go wrong. Twelve unlabelled strings are hard to present, so most sites number them or call one the main address, which manufactures a hierarchy that does not exist and gives an attacker a clear target: the top slot. That is the reason this site refuses to number or label them, which is covered on the mirror page.
The honest summary is that twelve is neither reassuring nor alarming. It is a larger set than most, it behaves statistically like twelve independent addresses, and the number itself is not evidence about anything. A market with one address and a market with twelve are equally capable of being anything at all.
| Set size | Reader can hold it | Displacement cost | Ranking pressure |
|---|---|---|---|
| One address | yes | lowest | none |
| Three addresses | roughly | moderate | low |
| Twelve addresses | no | highest | high |
The middle column is the one people underrate. A set small enough to be remembered is a set where a change gets noticed. Past that point the reader is trusting the publisher completely, and the size of the list stops helping them.
Questions
Is publishing twelve addresses a sign of a large operation?
Not by itself. Generating addresses is cheap and the number published is a choice about presentation as much as about infrastructure. No conclusion about scale follows from the count.
Does a bigger set mean better availability?
It could, but this site cannot say whether it does, because it never opens an address and has no measurement of any kind to offer.