Right, but you seemed to be arguing that illusory images meant that the curvature itself was illusory, whereas in my toy universe space really is toroidal, and anyone using the existence of multiple images to suggest otherwise would be wrong (because I set up the universe that way). In fact, such images would be convenient evidence that space was actually toroidal. I don't see how Einstein rings are any different- they can't be evidence against actual curvature of actual spacetime. You'd expect double and triple images if spacetime was curved.

One prediction of GR is that if there were enough mass in the universe, the global curvature would be positive everywhere and the universe would close in on itself and be connected more like a sphere than a plane. If you moved in a straight line you would eventually return to your starting point. That doesn't seem to be the universe we live in (as far as anyone can tell, it's flat) but that certainly seems like a prediction about the actual curvature of actual spacetime.