Eric is reputable...BUT...I can take out my Surf PI pro right now and show video that Eric statement isn't valid for all machines. If I'm lucky, I air test at 6". I'll nab coins at up to 3x that depth when they are buried in the ground.
I have heard this from numerous people but I have not been able to replicate it and I have tried many times. I have some theories why I can't replicate it though...I have tested a CTX, Excal, Whites DF, and CZ21. All air tested better than the ground test. In fact, the air tests were nearly double the ground test. That doesn't mean I believe air tests are always deeper...
To do my tests I first made sure I went to area free of emf. I have an emf tester that confirmed this. I have a pond/field behind my house that isn't near anything and could do it there. Then I air tested. Then I put the quarter into the ground, attached to a plastic ruler. The ground was a mixture of soil/sand as I live near the beach. My beaches on the US east coast have very low mineralization so you can get max depth with them (why most PI and VLF depth is usually the same here).
In my air tests, I would get 12-14''. When I tested the machines in the ground test, only the CTX could hit 8 inches and that was with a 17'' coil. The rest hit 6''. Now there could be problems with my test, in that the ground could of caused the issue. Also, the halo effect. So I did let the quarter in the ground for 3 days but nothing changed though that isn't really long at all.
I have spoken to OBN about this on another forum a few yrs back who said that with his headphone adjustments to the Excal his machines will ground test deeper, so that does confirm that machines can ground test deeper. I have heard that the BBS/FBS technology looks for ground, so an air test is not supposed to be valid on them. This is of course according to minelab though Tom Dankowski confirmed it as well. However, on my tests I cannot get them to ground test even close to the air test. My machines are stock though, not modified. So many things can affect these tests though - ground type, rust around the object, emf, swing speed, settings, etc.
So my conclusions: My test leads me to some conclusions but its far from a perfect test. I need to do a much more controlled test in the future. I plan to do this with the EQX but my hypothesis is that a machine isn't going to really be deeper on an air vs ground test, at least for a recent drop. I don't think you can really do a fair air test against a drop that is old and has tons of rust around it. That should be deeper then the air test. I think this is where we get a lot of mixed information from on air vs ground tests.
Anyway, its an interesting topic. I don't think we really do a lot of controlled tests. Most are eyeballed and we don't really reflect on them. This is one of the reasons I love seeing tests by eric foster, dankowski, OBN, etc.