Which explanation comes closest to how zipping a file shrinks it?

Compressing a file relies on spotting repeated patterns and rewriting them in a shorter notation — a run like 'AAAA' can be stored as 'A times 4' instead of four separate characters, and since real files are full of such repetition, this rewriting shrinks the file size while remaining fully reversible back to the original.

Files don't contain physical 'empty space' that gets squeezed out in any literal sense, and lossless compression specifically does not delete less important content — that description actually applies to lossy formats, a different category entirely.

JPG images and MP3 audio use lossy compression instead, deliberately discarding fine details that human eyes and ears barely notice, which shrinks files even further than lossless methods can but means the original data can never be perfectly recovered afterward.