More recently, I've heard other figures cited for the role of slide/frame fit in precision -- and 8%-10% seems common.
But isn't this really talking about Ransom Rest results, and not aimed fire? There may be something else at play here, and I'm exposing my ignorance -- certainly not the first time -- but I simply don't understand.
Back to the earlier point -- if you're using the sights on the slide, and the slide/frame (and sights) align consistently when the gun is cycled, why would slide-to-frame fit play even that big a role?
There are guns that have loose slide/frame fits but shoot very accurately. Most of the polymer-framed guns have reasonable slide/frame fit, but don't always come back to the exact same slide/frame locations, thanks to the give in the polymer. But some of them are very accurate.
That's why Ransom Rest tests are not used as much with polymer-framed guns as with metal-framed guns. Polymer frames affect the gun's precision -- its ability to return the entire assembly to the exact same place with each shot. A RR test measures a gun's precision, which is its ability to return it's part to the same place with each shot. The sights aren't used.
If you take one of the best 1911s, and put that top-end slide assembly on a frame that is more loose, I would expect sighted fire would be about the same for either frame, but would expect the Ransom Rest results to be much better for the slide and frame that were a matched pair.
It's only when you put the gun being evaluated in a Ransom Rest, it seems to me, that slide/frame fit becomes critical. Or, when the sighting system used is mounted on the frame. Then it could make a difference.