Right.
I think that it's not difficult to make a program that, given a single word, examine every varation of:
Adiacent char switch (ex: nmaepros.com, namerpos.com, ...)
Single char skip (ex: nameros.com, namepos.com, ...)
Single char doubled (ex: nammepros.com, nameprros.com, ...)
Single char wrong typed (ex: mamepros.com, bamepros.com, kamepros.com, ...)
..etc..
And then check various search engines for the original word ("namepros") frequency against the new typo frequency; generating a sorted list with the number of times the word is misspelled.
At the end, only the user brain could decide if a typo is less frequent because people tend to correct it, or if it is more frequent because that word means something.. etc etc.
But i'm sure that there are lots of great programs around there..