Here is an example of a simple 2f system to compute fiber coupling (located in Zemax samples).

The MFE operand used here is FICL to calculate the Fiber Coupling efficiency [η]. When using Zemax Tolerance, I set the Criterion: Merit Function → Run Monte Carlo.

As you can see, the compensated parameter values are totally out of whack from the bounds I set.

Zemax documentation clearly states that “This bounding of the compensators is ignored if "merit function" or "user script" is selected as the tolerance criterion,  although the compensators will still be used.”

If that is the case, then what is the right way to bound the compensator(s) when using a Merit Function Criterion?

Nominal System

Perturbed Monte Carlo system with compensation: The compensation is totally out of whack.