Another way to confirm would be to measure, then rotate the car 180 degrees and re-measure. Level would be halfway between the two measurements.Yeah, the risk comes if the ground isn’t as level as I think. However, I have tried parking my car in different areas of my garage (3-car garage) and gotten the same result, so I know it is close.
But yes, even a 3mm surface undulation, which would be difficult to detect, would result in about a half-degree error on actual camber. So, while this method does require the ground around the wheel to be pretty smooth the flip side is this method does not require the ground to be perfectly level, which is probably even less common.
The other home camber method would be using a good ($100+) bubble camber gauge arguably capable of <.25 degree resolution and precision, but it would be difficult to confirm that you haven’t inadvertently parked on a 0.25-degree slope, which would also drive a half degree of total camber error.