Seraina Hartmann

If I see what you are asking to search on, it includes ?
Those segments where you have 100% also include the ?
Where as the 95 / 90 do not.

The search / hit logic when running concordance is different to when you are doing a TM lookup so that explains why when I see your 95%, 90% etc they vary a bit.

Lyds

Hi Lyds

Thank you very much for your reply.

Your explanation refering to the question mark makes perfectly sense.

Unfortunately, I still don't understand on what the score is based on. I might have to rephrase my question.

I would like to understand what the 90/95/80% refer to?

For instance, the score in a fuzzy match refers to the source segment retrieved from the TM compared to the new source segment in the editor. But what does the score in the conocrdance search refer to?

Thank you very much for your help.

Best

Seraina

Update on this thread with the answer I got from Trados Support (for anyone interested):

"The %-match is based on the searched term and the results from Translation Memory.

In first step it makes the fuzzy search which is searching similar terms depending on your search.

In second step is making a scoring on the segment depending on how accurate/similar the terms are depending on the information from TM."

So, my confirmed conclusion was:

• The scores including the question mark result in 100%.
• The 95% result includes the exact sequence of the searched expression including question mark, but also some additonal characters (namely “keit” in “Erwerbstätigkeit”). This is why it is not scored as 100% but as 95%.
• The 90% results miss the question mark. This is why the score in this match is lower (90%) compared to the one that includes the question mark (95%)

Best

Seraina