For anagram scoring, why not just determine the overlap between both words in terms of shared bigrams (i.e., two-letter substrings), relative to their length? H

For anagram scoring, why not just determine the overlap between both words in terms of shared bigrams (i.e., two-letter substrings), relative to their length? Has anyone tried that? Is it equivalent to what mjd did?