I think this is the key. Ingress was a successful AR game, and the geolocation was the only AR, no shoddy camera overlay. I think you can draw a parallel to traditional video games, in that games with a general understanding of maps, and not much small-scale definition(MUDs, Civilisation, even 8-bit platforms with no sub-pixel movement) tended to achieve a greater level of achieved complexity by glossing over small details, in the same way that Ingress/Pokemon Go only placed the locations in a general geographic area.