Which is somewhat ironic, because I know that "sit and just do it" is the ideal start. Think less about choosing, just pick it up and go.
But I end up spending more time looking up things and getting excited about them.
LOVE2D? Oooh, it's lua! It can even run on Android! DragonRuby? Oooh, simple and lightweight and fast! Oh wait, Godot just had a new update! Hey, maybe I should get back to programming Java, libGDX looks neat! Damn, Raylib feels like it does many things right! And on it goes...
I overwhelm myself with choices, start a bit then abandon at the second hard-ish hurdle (like menu/interface showing under the map despite lots of fiddling with the Z position, in Godot). So, yeah, just exposing my problem, I suspect I'm not alone in this.
I think it's fine to engine hop a bit as long as you actually build something interesting in each one. Even if it's the same game every time, you're still learning different ways to do something. That's not a waste.
If your goal is to finish a game, you should maybe take notes about the engines you've tried and set yourself a deadline for choosing your favorite.