It definitely is related to prediction by partial match (PPM). BWT sorts rotated data and what is achieved is that same suffixes group together: ... "Bzip

It definitely is related to prediction by partial match (PPM). BWT sorts rotated data and what is achieved is that same suffixes group together: ... "Bzip2 and Bzip3 are simply combining more" "Bzip3 are simply combining moreBzip2 and " The preceding (to suffix) character goes to end and then gets outputted. This is much like PPM going backward. There is a PPM* algorithm (unbounded context length) where authors considered reconstruction of contexts from data, utilizing something like LZSS seach. Same idea is in BWT - context is reconstructed from data. BWT also breaks near dependencies in data, this is why move-to-front with Huffman or arithmetic encoding works well there.