As said above, byte comparison would be considered the most exact as it compares each file "byte-by-byte". The other methods use hash fingerprinting. MD5 has the highest chance of a "collision"
Because "byte comparison" is much faster than MD5, Why not add "byte comparison" in [Duplicate File Finder] ?
[Duplicate Cleaner] has "byte-to-byte", md5 ,sha1,and some other hash values in [content comparison].
Byte comparison would not always be faster and would be orders of magnitude slower when the are lots of files of the same sizes. They’d all have to be compared with each other in every combination, instead of calculating each file’s hash once.
Hash collisions are unlikely to happen unless someone is creating them on purpose, which seems an unlikely attack against a duplicate finder tool that you run on your own files.