I had a little play with the maths. A Lancaster at a height of 20,000 feet en route to Hamburg would lose a G-signal from Stenigot around coast in of Holland. A Mosquito at 30,000 feet gained a further 39 miles and would lose the signal near Groningen. To get a G-line you need a signal from two stations and for a fix, 3 stations. The Dutch coast would be the absolute maximum distance.
From the Dutch coast the shortest distance would be another 156 miles for the Lancaster or around 40 minutes. If we assume a DR error of 10 nm/hr, plus an error of 2 miles on the last fix, you have an error of 8-9 miles. That can be reduced by getting a pinpoint but that would still have an error of 2 miles assuming an accurate identification.
Wind finding was pretty primitive until the advent of Doppler or a high definition radar system.
The advent of H2S would have been of huge benefit, once they were permitted to use radar over Germany. By 1944, 35 Sqn (a pathfinder sqn) F540 recorded bombing errors in the order of 400 yards. It took time to get things together.