... just the same way as multi-volume archives do. Say Trados could skip 100% matches and repetitions for calculation purposes and then equally divide all other segments (fuzzy and brand new) by any requested number. It would be even better being able to tick the matches that I want to exclude/include from/to such splitting process. For example:
Divide by [3] based on:
_ Reps - 12 words
_ 100% - 360 words
_ 95-99%- 420 words
v 85-94% - 100 words
v 75-84% - 150 words
v 50-74% -2000 words
v NEW - 3909 words
where _ stands for empty checkbox, and v stands for ticked checkbox
Then... 100+150+2000+3909=6159 words / 3 [translators/portions] = 2053 words each portion