As a user, I expect thresholds with the same names applied to different sides of a pairing to be graphed correctly

[james-d-brown]

Given an evaluation with the same thresholds to be applied, separately, to each side of a pairing (observed and predicted)
When that evaluation includes graphics
Then the graphics should properly distinguish the different thresholds

[james-d-brown]

An example to the contrary in Redmine ticket 156712.

[epag]

Dont believe this was completed, moving to 7.6