At the very least I am aware of where the front sight is relative to the notch and I know how precise the sight picture has to be to break the shot.
It's been a while since I lurked or posted on brianenos.com, but I see some foundational elements here that apply to the great pointshooting debate.
1. The first fundamental is a good Natural Point of Aim. This is quite flexible (within limits) with handguns, since the hands do legitimately tend to follow the eyes. Too much flex and you will wind up forcing the firearm to NOT return to its starting position, and analyzed by Lurper.
2. From NPA comes the "index", an ability to *know* where the pistol is pointed because of your body awareness. I know muscles have no "memory", but a lot of people call it that--just like how you *tend to* reach for the shift knob at the same place in every car, based on your primary daily driver's layout. Your muscles have nerves, and the memory lies in your head, guiding the muscles.
When done as a Lurper two-sight-picture "double tap" or "controlled pair" with a separate sight picture for every shot, I'm not sure that there is truly a DECISION to fire the 2nd shot only AFTER seeing the sight picture.... In any case, the awareness quoted above is comparable with the IPSC shooters' "see what you need to see to make the shot". I have a general idea of how far out from the center of the rear sight notch my FS can be at 7 yards, and still get a hit.
Now, I've tried Mr. Temkin's general ideas from as far back as about 1995, before I ever read about them, and did visual index looking past the twin images of the slide while looking at a 10- to 12-yard metal target only 8 inches wide. I fired as quickly as I could get that time of reference, in dusk too dark to see the whitedot sights, and got hits every time. Estimated splits were maybe twice those in Lurper's video.
In my opinion, it's still sighted fire as long as you are using some visual reference to the firearm. You are simply using a different object to "sight" on, even if it's the whole slide. Coarse sighting is still sighting, and has proven to be VERY effective and fast.
Lurper: repeat your exercise from low or medium ready, just with your eyes closed before you raise the gun. NO sight picture at all, and I'll betcha your shots are still within 4 inches of each other, though they might not be quite so close to the center of the A Zone.
Mr. Temkin: Try your techniques under circumstances where you cannot get any visual reference to the gun or its slide and generally cannot see it at all. I postulate that the most well-trained in your sightless techniques will be almost, if not just as, fast AND accurate as when they can see the gun, out to at least 5 or 7 yards. Even those casually trained for only a half-hour will probably get acceptable speed and accuracy to 5 yards.
I base this on some of my own PS experiments with a coupla snubbie revolvers in the late 1970s. One-hand HIPshooting worked okay maybe half the time out to 15 yards (with the occasional memorable aberration center hit to 25 yards!) only when I could see the gun. Tried it with a piece of paper held at my neck, obscuring the position and orientation of the gun, and hipshooting fell apart and was marginally good out to only 3 yards.
Bottom line: the "see the sights every time" factor is sometimes more of an incidental observation of the sufficient alignment for each shot, which will happen regardless of where your eyes are focused if you have a good NPA and Index, including sufficient awareness of where the slide of the gun is pointed in relation to the target.
Few PS techniques are what I would call true PS shots. You're still SEEING what you need to see to make the shot.
IMNSHO, the true key to second-shot accuracy is letting the gun snap back after recoil, and triggering the shot only when it gets back to the NPA. 7 yards and closer, you might not need any visual index, but it will surely help for the final timing and closer placement of follow-up shots.