A lot of experimental and applied physics operates this way. If you are synthesizing material, for example, it takes a lot of time and effort to get high yields of what you want. Before that your processes can be very probabilistic.

In fact, just finished listening to a talk where a experimentalist was talking about how to get the fabrication yields of superconducting qubits from currently low double digit to 99.99+.

Every scientist does that at some point. I've easily crossed my fingers and hoped numerous times that code I'd written would work, especially on the first time. Even more rewarding in the superstition when the project is hard, and you're a bit daffy at the end.

It's a human thing.

Surely Feynman made jested comments before running experiments. I'm sure some digging in his wonderful books and letters will find many examples.