I would say this is incorrect.
Measuring SWR with the load disconnected mainly tells you how mismatched the tuner output is, not how much power is being lost inside the tuner. Assuming that no power is reflected when the SWR is adjusted to 1 ignores internal losses, since some power can be dissipated before it ever returns to the measurement point.
A port-to-port (through) measurement gives the total loss, which includes both mismatch losses and internal (efficiency) losses. By combining this with a reflection measurement that quantifies mismatch, the mismatch contribution can be separated, allowing the remaining loss to be attributed to the tuner’s efficiency.