Predicted best AFMA -534

I'm trying to do some AFMA on my lenses and so have downloaded FoCal for OS X. When I run this, I have to use manual mode which involves taking a series of pictures (3 sets of 3 from -5, 0 and +5).

When I get the tool to analyse the files it predicts AFMA at -534 on my 35L on a 5D Mark iii.

Would this indicate a bad copy of this lens? It doesn't have a filter on it.

Think I may have solved it (partially). I had the lens set on f11. When i dialled it down to 1.4 and did the same test, it came back with prediction of 2. I did the test again with AFMA set to -2, 0 and 2 and it predicted 1. I set the focus point to just to the left off the centre point and did the test again. It predicted 4. Confused. I changed the aperture to 2.8 and did the test again, it predicted -4.

Is this normal behaviour?

2 options:
-Something went wrong with the test, try again
- the lens is front or backfocussing a looooooooooooooot

No, it most likely means something is not right with your setup. FoCal analysis can only be as good as the quality of the data you are feeding in. There are many factors which can make FoCal analysis unreliable, I'll just mention a few (read their guide for the complete picture):

- Is the light bright enough? I think you need at least 10 EV, better still 11 EV;
- What is your light? If it is fluorescent or LED, you might get very bad data. Halogen would be better; but the best is daylight.
- Did things change during the shooting? Did the light change brightness, did you move the tripod accidently?
- Are you at the right distance from the target? It should be 50x FL or more.

I was using a flash gun, no lights on in the room but there was some daylight. Would the flash be inconsistent? I set it to manual and had a good exposure each time.

Tried my 70-200 at 200mm and 2.8. It originally said 2 and then when I did a test between -2, 0 and 2 it predicted 0. I wonder if this is not the best way to test. Unfortunately there is no automatic way of doing this and my eyes are going cross eyed from trying to see differences in focus from black and white print on a computer screen. It's not easy.

were you able to even notice if the lens was off to begin with?

It seemed to be sharp on the centre focus point and close up, even at 1.4. But I've had some landscape shots look a out of focus that were using an outer focus point.