GG, possibly quite right. I have an article at home in the Journal of Navigation about the effects of GPS jamming and a comparison of the GPS apparent track with the eLoran track in the jammed sector. The article also illustrates where some ship's courses were overland.
One figure shows GPS errors in the region of 8.1m but it is not clear whether this is miles or meters. Given some of the other illustrations it might well be miles.
I haven't absorbed the full article but it would seem to suggest that a GPs-eLoran combination is a potential solution.