It works for me ...
Sure, it's
usable w/in the limits it creates (which are significant IMO), but have you critically tested/compared the differences? I have, numerous times (although I don't own Imatest).
For these images/uses I can fairly well promise you that you would be better off just cropping; as long as there was at least 1024x pixels remaining, and maybe even if there wasn't. Both in terms of IQ and functionality.
Also, being at f/5.6 max on the D850 allows you to stop down
a little for sharpness/DOF w/o a significant loss; but being at f/8 max doesn't (diffraction limited)... if you are getting better results at f/10 then that means the TC is compromising the native lens/sensor resolution beyond where it can be recovered.
According to
this test, the Imatest results for the 200-500 show it delivers 2,244 L/PH at 500/8 (center), a bit better than the 2,143 L/PH at f/5.6. With the 1.4vIII it's down to 1,620 L/PH at 700/8... it does show an improvement to 1,894 L/PH at f/11, but still below the native lens even wide open (f/5.6). With the vII TC I believe the results would be even somewhat less for that lens. Basically, you're just smearing the details out in order to make them larger on the sensor... and compressing them back down to 1024x isn't really going to help comparatively.
It's worth noting that the test was done with a D4s... on a D850 I would expect a little higher baseline numbers and less improvement from stopping down. Basically, I think the test results would look a lot like shifting all of these 1 stop left.
If you're happy enough with the results using it, then that's all that really matters. But I really doubt that it's really helping you at all...