What you can do with one photon is to observe that it almost never appears in the screen in positions where geometry dictates there would be destructive interfe

What you can do with one photon is to observe that it almost never appears in the screen in positions where geometry dictates there would be destructive interference of the photon was a wave. If you literally try one photon only once, well, you can't really talk about probabilities; whatever happened happened (possibly you witnessed an unlikely event). But if you repeat the experiment N times, even if you do it spaced by many miles and many days, and collect the data you'll notice that the measurements cluster in ways that are compatible with interference bands. If an actual physicist can tell me whether that I wrote above is fundamentally wrong, I'd be happy and have more questions