Technically speaking…
Let’s say you have a destination with 6 files {a,b,c,d,e,f}.
If you are copying {a} to the destination then without no-traverse, rclone will load in the definitions for all the files {a,b,c,d,e,f} before discovering whether {a} needs to be uploaded. If you use --no-traverse rclone will just check {a} on the remote.
So why wouldn’t you use --no-traverse all the time?
If you are copying {a,b,c,d,e,f} to the destination, then rclone will check each file individually. This will take at least 6 transactions, wheras likely you could have got the listing for all the objects done in 1 listing.
So there are tradeoffs! The new sync method implemented in 1.36 makes --no-traverse less useful than it used to be, but it can still come in handy, especially if you are moving or copying files into a deep hierarchy.