You are correct; I confused archiving with compression. However, even considering only the compression process, same compression parameters cannot be guaranteed, as it is unknown which compression parameters the image publisher used.

That is not correct. You would have to use the same compression tool (and likely version) for this to match.

Old docker discarded the compressed bits but kept some metadata about the the so it can at least recreate the tar.

It also recreated the manifest o push.

Zstd for example only promises determinism on the same version of the library. I've personally seen the hashes mutate between pull and export. Things like tar padding also make a difference. Really, the thing to do is to hash on the _uncompressed_ data and let compression be a transport/registry detail. That's what I've done, at least.

The whole entire reason is compression is not deterministic across tooling.