I've never really understood why the bother of calculating focal depth at all when all you have to do is adjust it until you get the best SQ. If the best SQ happens to be achieved at a level other than the calculated value I doubt we'd opt to use the calculated level instead.
Is it something some people do just for a point of reference?
Well, it may depend quite a bit on the specific dish/feed you have, but in my case, it is extremely difficult to adjust the focal length. With my dish, it's virtually impossible to move the feed in/out while watching S/Q. I have tried dozens of times over the past 14 years since I got this darn dish, and every time I have given up, and gone back to setting the FL according to the calculations. Basically, with the quad feedhorn support arms, to change the FL, you have to get to where each arm connects to the dish, and change the length of that arm, keeping each arm the same. This is hard enough that I never do it let alone the idea that it can't be adjusted real time, but rather set it, then check the S/Q. An alternative way of setting FL with my dish is to move the feed throat in and out via the scalar ring, however if you do this, you're changing the F/D setting for the feedhorn, plus, when you loosen the allen screw to do this adjustment, the throat cocks one way or the other plus you have to be careful not to change the polarity setting, plus it can't be done without standing directly in the way of the signal, so if I move at all, it affects the signal. So the signal is bouncing all around when I try this.
There is another way that I've tried adjusting the FL, and that is by twisting the whole feedhorn/scalar, so that the quad arms aren't aiming at the center, but are more tangential. This pulls the whole feed in closer to the dish changing the focal distance without affecting the F/D setting of the feed. The only problem with this is that doing this changes the polarity, so that I can't really rely on real time adjustments while moving the feed in and out, but instead have to make a change, then go back inside, adjust the polarity via the receiver, check to see if it improved things, go back outside, and try again, and again.
But perhaps the biggest issue with respect to adjusting the FL by peaking S/Q is that if you do this on C-band, the changes aren't significant enough to observe the S/Q changing more than the normal variations, so it's more accurate to peak on Ku. The problem with peaking FL on Ku is that if your dish is even very slightly off aim, you will get maximum S/Q at the wrong FL, because the signal pattern becomes more of a donut shape when you're at the wrong FL, and if you peak the signal when off FL you'll peak on the side lobe, and both the FL and position will be peaked at the wrong place. Ie basically, you need to be changing both the aim of the dish and the FL at the same time, which with a dish controlled by a receiver inside the house makes any real time adjustment extremely difficult. I've basically given up the idea of peaking S/Q, and just set the FL to the calculated values. However usually about once a year, I get all enthused and try again..... but each time I give up and set it to specs.
NOW, if I had the button-hook feed I had on my first dish, I think I could peak the FL. But not with this quad feed arm arrangement. Anyway, it's not something I "do just for a point of reference".
I just can't figure out a way to do it that's better than using the calculated values.