Abstract: A lossless data compression method that enables fast decompression is presented. To roughly characterize the method essential is that the finding of the repetitions is based on symbol pairs. The method repeatedly finds pairs formed by two adjacent symbols and replaces occurrences of common pairs with one new symbol. Similar prior art methods store the replacement rules into a separate entity that itself has to be compressed using some technique. The presented method does not need the replacement rules—the compressed symbol sequence is built in a new way, which enables that the replacement rules can be inferred from the symbol sequence. In prior art decompression, the replacement rules are decompressed and then consulted to interpret the symbol sequence and produce the original uncompressed input. The presented method enables decompression without recursion in one sequential traversal of the compressed input.