then the wordcount is a little bit of a moving target depending on Trados versions, penalties or the exact position of the Earth in the Galaxy at the time the word count was done. It's amazing that people actually get paid out of those word counts!
Normally our Trados word counts never match with the client's to the word, so the rule of thumb we follow here is just to go ahead with the client's word count if it is acceptable compared to fighting it or troubleshooting it. After all, the cost per word is just an arbitrary value that you set and you can factor the client's behaviour in your rate after some experience with him or her, for example if you detect that he turns off all the penalties to reduce the nominal cost.
I'd say that the most significant sources of variance in order of importance are:
2. Trados version
3. Word counts done with .rtf files and you are handed ttx files
4. Position of the Earth
Natalya Zelikova wrote:
I have received a project from my client and after I analysed the files against empty TM I found that the total wordcount is slightly bigger than given by the client.
The total wordcount was only few words more than client's one, but distribution of words between "no match" and "repetitions" was also different.
The client has replied that the reason for this may be different TM settings, and though this time the difference was insignificant, I'd like to know if there are any "required" TM settings for accurate analysis?