Then why hire salaried workers while tracking work via sprints?
>your velocity might be half or less what it could be if everyone ended up on-site.
This is assuming more work actually gets done in an office and that offices don't have their own set of distractions. People can also slack off in an office too, or become so distracted by others that their velocity suffers.
Measuring employee performance seems to be a continual struggle that has no relation to in-office vs remote workers. If you can't accurately measure success & value generation of remote workers, you probably also can't measure office worker's either. Just checking that someone is physically in their seat doesn't actually mean they're adding tangible value.