800-541-8270
A “Fuzzy Match” isn’t a fire hazard; it’s an indication of how well SDL TradosTM recognizes a text segment in a document.
Translation Memory (TM) programs look for segments in memory…
that have enough content in common with the current document segment for the translation memory to be of use for translation. The match is expressed as a percentage. A complete match is called, not surprisingly, a 100% Match.
Less than 100% is a Fuzzy Match. In general, 65-70% is considered the “fuzzy match threshold.” Depending on the settings chosen by the translation manager, the threshold can be as strict as 99%, or s flexible as 50%.
Any matches that fall below the threshold are considered No Matches. For text segments that cannot be found in the TM, Trados marks the first instance as a No Match, and every other instance as a Repetition. This means a new translation can be entered once, then propagate to all instances of the segment.
It is actually possible to have a match that is BETTER than 100%. This is known as a Context Match.
To qualify as a Context Match, the translation memory segment must exactly match the document segment, and both must be preceded by an identical text segment, as well. And yes, there is a match that is even better that that: a Perfect Match matches segments on BOTH sides of the segment. Who knew 100% only gets you third place in Trados?
Remember: the more translation that is taken from the TM, the less time spent by human translators, and the lower cost to you.
In general, a segment corresponds to a sentence; it is a string of text that ends in a full stop. Again, context is key to what makes TMs so powerful: they don’t just match word-for-word or string-for-string, but frame the words within the context of how they are used to prevent mistaken matches. For example, the phrase, “I saw him by the bank” could refer to a financial institution, the side of a river, or a bend in the road. Knowing the whole story requires the larger context that TMs record.
The following examples contain multiple language segments using segment B as the match type:
| To Match or Not to Match? It’s a question. | |||
| Match | Document Segment | Translation Memory Segment | Explanation |
| Perfect Match | A.) HAMLET B.) To be, or not to be, that is the question, C.) Whether ’tis nobler in the mind to suffer The slings and arrows of outrageous fortune, |
A.) HAMLET B.) To be, or not to be, that is the question, C.) Whether ’tis nobler in the mind to suffer The slings and arrows of outrageous fortune, |
In a perfect match, segments A, B and C are identical in the Document and Translation Memory segments. |
| Context Match | A.) HAMLET B.) To be, or not to be, that is the question, |
A.) HAMLET B.) To be, or not to be, that is the question, |
In a context match, segments A and B are identical in the Document and Translation Memory segments |
| 100 Percent Match
(Complete Match) |
B.) To be, or not to be, that is the question | B.) To be, or not to be, that is the question | In a 100 percent match, the B segment is an identical match. |
| Fuzzy Match | B.) To be, or not to be, that is the question |
B.) To bee, or not to bee, that is the question |
In a fuzzy match, some, but not all, of the words match in the segments. |
| No Match | B.) To be, or not to be, that is the question |
B.) This above all, to thine own self be true. | In a “no match” scenario, two thirds or less of the words in the segment match. The given example is a 10 percent match. |