I have a source file with 3 segments:

By the second grading period following the, of different species that help them live and survive, such as webbed feet in aquatic animals, with no more than 3 verbal prompts with 70% accuracy on 5 assignments per grading period.
By the second grading period following the, of different species that help them live, such as webbed feet in aquatic animals, with no more than 3 verbal prompts with 70% accuracy on 5 assignments per grading period.
By the second grading period following the, of different species that help them live, such as webbed animals, with no more than 3 verbal prompts with 70% accuracy on 5 assignments per grading period.

The 1st is a modified segment I already translated and the 2nd and 3rd are a further modification.

Once I Prepare the file, the 3 segments get populated with a 83 %, 80 % and 73 % match respectively.

When I translate the 1st segment and I move to the 2nd one, instead of getting the match from the 1st segment I get the old match (80 % from the TM)

But If I close and open it I get the proper TM match of 97 %

Does anyone know why it is that I have to "close" the file in order to get the TM to show me the "Best" result?

The TM is set to update.

Any ideas?