It may be useful for a repository to count as more than 1 copy of a file.

Examples could include: A repository on S3, when a bucket is configured to be sufficiently durable. A --sameas special remote that accesses data copied from the other remote to another storage media (eg by a backup process). A repository on a RAID array.

If this were made configurable, it might also make sense to allow a repository to count as less than 1 copy, eg 0.5 copies.