@eobox91103 thanks for the award 😂
About your findings:
Windows itself does not “fence off” or reserve disk space based on a file's self reported size. Most torrent clients do not pre-allocate by default. Instead, they create sparse files: files that have a full logical size (e.g. 30 GB) but only consume disk space for the pieces that have actually been downloaded.
This means nothing prevents you from starting two 50 GB torrents with only 50 GB free, both will download simultaneously until the disk fills up, at which point they will fail.
For example:
You have 100 GB free
You start downloading a 10 GB file
At 50% completion, only ~5 GB of disk space has been allocated
You still have ~95 GB free
Even though half the file does not exist on disk yet, when reading the file, NTFS returns zeroes for the missing regions, as most user-level applications cannot observe “holes” in files, including Windows Explorer. There are backup tools that can copy sparse files exactly, without zero-filling the sparse regions.
When you copy such a sparse file using Windows Explorer, the copy process reads the entire logical file:
Real data for downloaded pieces
Zeroes for sparse regions
Those zeroes are then written to the destination file, causing full physical allocation. The copy is no longer sparse and now consumes the full disk space.
So when you say that copying folders to backup drives doesn’t waste space, that’s true for completed files. But copying an incomplete sparse file expands the holes and wastes space by filling them with zeroes.
In normal backup workflows, this is not a concern because completed files already have their full data written. Logical size and size on disk will naturally be very close. For complete files, the Size on Disk will usually be greater than the file size itself, things like the file name and path length can increase the size on disk by a few bytes, for example.
So in short, I hope you're not backing up incomplete files, and if you're not you're all good 😅
