Measuring PV string similarity via Levenshtein Distance?

[benstoker]

Could the Levenshtein Distance algorithm be used as a metric to determine the similarity of two PV strings?

Example:

PV1: "Qe4+ Kg8 Qxb6 Nd6 b5 axb5 cxb5 Nd8"
PV2: "Qd4+ Kg8 Qxb6 Nc6 b5 axb5 cxb5 Nd7"

Computed LD is 3. An exact match would be 0.

The point would be to run LD comps on 2 sets of PV strings and thus get a metric of the similarity of the two sets of a large series of PV strings.

See http://en.wikipedia.org/wiki/Levenshtein_distance

[Dann Corbit]

Sometimes it might be a good measure, and other times a single character changed in the pv will be the difference between getting checkmated by the opponent and checkmating the opponent.

What is it that you hope to achieve by this calculation?