Science is a process for learning knowledge, not a set of known facts (or theories/conjectures/hypotheses/etc.).
Phlogiston theory was science. But ultimately it fell apart when the observations made it untenable.
A belief in luminiferous aether was also science. It was disproved over time, and it took decades from the Michelson-Morley experiment to design robust enough studies and experiments to prove that the speed of light was the same regardless of Earth's relative velocity.
Plate tectonics wasn't widely accepted until we had the tools to measure continental drift.
So merely believing in something not provable doesn't make something not science. No, science has a bunch of unknowns at any given time, and testing different ideas can be difficult to actually do.
Hell, there are a lot of mathematical conjectures that are believed to be true but not proven. Might never be proven, either. But mathematics is still a rational, scientific discipline.
Ok, thanks, I'll just set a threshold for myself. If I do more than 6 reps at 95% for my 1+ set, I'll do a joker set for 105%. If I bang out at least 5 reps on that first joker set I'll do 115% for a second joker, before going into accessory work.
We'll see how that goes.