I'm very thankful for Toldo's suggestions on text compression and am wondering about general-purpose compression. Right now, I have several versions. A special version is doing exceptional, but my main ones are doing horrible on many files. I'm currently using LZ77, Adaptive Huffman, Elias2, my Last16 technique, a method to offset from the last two LZ77 blocks and a way of compressing AdaptHuff literals that are uncompressible by counting other uncompressible values and shortening them to just enough bits to uniquely identifying them. Does anybody here have any ideas on how to better this?