43", my experience over the past 25+ yrs does not align with your comments. Not that I am suggesting that you are incorrect, just that my personal experience with Sony Pixus, Garmin Pilot III, Garmin 296/396/496, Garmin 250XL, Garmin 430W/530W, flying all over Oz, differs. It has always amazed me just how accurate these things are. I have only seen the G430 in the Bonanza go into DR mode once for a few minutes in 800 hrs, and I currently have 10 satellites on the Humminbird chartplotter sitting on my office desk in Townsville, without it having a clear view of the sky. The GPS tracker in my boat is hidden under the deck and can't 'see' the sky at all - but it still sends its position accurate to a couple of meters.
I fly for a living using GPS so a lot of hours navigating by it, while its 99.9% foolproof I've had several missed approaches due to no RAIM at the FAF or past it, and several events where the GPS navigation solution became suspect and was confirmed off track. Some of the events were unit related and others satellite or atmospheric related. Ice on the GPS antenna can affect it as well as moisture in the wrong places. It is not regular, but it does happen. This is where installation type really can make a difference. The key being what information is displayed if the navigation unit is operating in a degraded mode, and when will it tell you so. For instance in a TSO unit will not allow operation in approach mode without RAIM, so it will dump flight guidance and make it clear you don't have the required fidelity.
This talk about trusting the unit when low among hills without having RAIM protection shows a lack of understanding of the drawbacks of GPS, like any other navaid.
The iPad as far as I know does not tell you if it has good/bad or no GPS signal.
This is also the issue with creating your own approaches in IMC, without the required programming it will most likely fly in terminal or en-route mode, which means different tolerances, alerting and RAIM requirements.