this is only a slight improvement (Score 1)

lzip is pretty good in those cases where it reduces the data to 0 percent of the original, but it is only 1 bit better than the following algorithm which has been around for years. Count the number of '1' bits in your data. If it is 1 return 1, otherwise call the algorithm recursively on this number. It is easy to prove that this algorithm compresses all data sets to a single bit in order log n recusive calls, where n is the number bits in the data. Is lzip this fast?