Matches are search hits in a translation memory. If a previously translated similar or identical segment is found in the translation memory for a segment to be translated, the search hit is considered a match.
Each match has a match rate. This is a percentage indicating the degree of similarity between the segment to be translated and the search hit found in the translation memory. The percentage of similarity is calculated on the basis of the content, i.e. text similarity.
The resulting match rate may be reduced due to other differences, e.g. different formatting.
In general, distinction is made between the following types of matches:
Search hit in the translation memory with a match rate of 100%. It fully matches the segment to be translated.
Search hit in the translation memory that does not fully match the segment to be translated. By default, the match rate of a fuzzy match in Across is between 50 and 99%. This may be due to different content and/or other differences, e.g. different formatting.
The minimum threshold for the fuzzy match can be adapted under Tools > User Settings > crossDesk > crossTank.
No search hit has been found for the current segment in the translation memory or the degree of similarity is below the defined match rate.
Apart from the three kinds of matches mentioned above, there are also other kinds of matches, like context matches and structure matches.