Translation Memory Match Calculator

Yes, across the full range of values this calculator supports -- more exactly matched words directly increases the effective match (matched words minus edit distance), which raises the match percentage, whether or not deletions, substitutions, or insertions are also present. The relationship can flatten at the 0% floor or the 100% ceiling, but it never reverses.

Match percentage is the effective match divided by the total source segment length -- the same number of matched and edited words represents a SMALLER share of a longer segment than a shorter one, so stretching Source Segment Length while holding matched words and edits fixed always lowers or holds flat the resulting percentage, never raises it.

A substitution is a word replaced with a different word (a name, date, or number change is the classic case), an insertion is a new word present in the current source that was not in the stored TM entry, and a deletion is a word that appeared in the TM entry but is missing from the current source. All three count equally toward edit distance in this calculator, regardless of which type they are.

Leverage Rate applies a discount by match tier because a fuzzy match still requires a translator's editing attention even when most words are already correct -- a 95-99% high-fuzzy match is discounted to about 90% of its raw percentage, a 75-94% fuzzy match to about 70%, and anything below 75% is discounted far more heavily, reflecting that low matches offer little real reuse value despite a nonzero raw percentage.

No -- this calculator clamps the result to the 0-100% range regardless of the raw edit-distance arithmetic. If matched words minus edit distance produces a negative effective match (heavy edits relative to matches), the percentage floors at 0% rather than going negative.