Jesse Good

The behaviour here is incorrect in my opinion too.  I also agree with your conclusion.  I'll raise this for discussion internally too and see what comes of that.

Jesse Good

... although while I was writing up the steps to reproduce this I started to have second thoughts :-)

The analysis is designed to show you the effort required to translate the file based on what will be made available to you in the TM.  So on this basis what is the point in having an analysis that shows you what you could get if you lowered the minimum match value when you work?  If you want to do that then lower the value before you run the analysis.

I actually think it's working as it should and it also makes sense.

What do you think?

Jesse Good

... although while I was writing up the steps to reproduce this I started to have second thoughts :-)

The analysis is designed to show you the effort required to translate the file based on what will be made available to you in the TM.  So on this basis what is the point in having an analysis that shows you what you could get if you lowered the minimum match value when you work?  If you want to do that then lower the value before you run the analysis.

I actually think it's working as it should and it also makes sense.

What do you think?

Paul

I see your point and if the analysis doesn't match with the actual results you get from the TM, it would be more difficult to judge the required work. I think the only issue is that the analysis says "50-74%" when in fact that it is showing you the word count for "70-74%". If you create your translation quotes based on the analysis, you would be billing the client for "50-69%" matches at the same rate as "New"...

Jesse Good

Jesse Good said:

I think the only issue is that the analysis says "50-74%" when in fact that it is showing you the word count for "70-74%". If you create your translation quotes based on the analysis, you would be billing the client for "50-69%" matches at the same rate as "New"...

I think this is just down to your own configuration.  If you wish to work with 70% in the analysis and then take advantage of a lower fuzzy when you work then you are not comparing apples with apples.  Surely you should change the rate bands yourself so that they say 70-74% and remove the lower values altogether?

I take your point though seeing as the defaults include a lower fuzzy but actually cap the segments that get in there by setting the search to 70%.

Funny I've never seen anyone picking this up before.

Jesse Good said:

If you create your translation quotes based on the analysis, you would be billing the client for "50-69%" matches at the same rate as "New"...

I would never allow matches lower than 75% to be weighted in any way. Since I know CAT tools, 75% was the lowest value for weighing. Even in times of Trados 3 to Trados 2007, when a 30% match could be similar to what we now have in partial matches, I always charged 100% for matches lower than 75%. This has changed nowadays, so that I try to get fully paid for matches up to 84%. This reflects the real amount of work needed to adapt such low matches.

So from my point of view there is no difference

Hi Jerzy Czopik

That would be one way of solving the problem

Jesse Good said:

That would be one way of solving the problem

Well, IMO the best way