what is the morphology transformation that justifies this "match"? I can see thinking that the suffix is -es rather than -s but ...
"Match all word forms" uses an algorithmic stemmer, not a dictionary-based approach; therefore, there will be both false positives (overstemming) and false negatives (understemming) in its output.
If you're interested in the rules, you can look here (although the site seems to be down currently): http://snowball.tartarus.org/algorithms/english/stemmer.html
