To keep your "level up" analogy going...
The first few levels of a startup are EASY. You're GOOD at this stuff. Ideation, deploying prototypes, early design work, etc. You're GOOD at it. You blow through these levels like a pro and you get the emotional reward.
Then you hit the wall where you SUCK. It's working, but not as well as you've hoped. There are more questions than answers. The next level is a long ways away, so you don't get the "rush of success" for a long time. You're not even sure you can COMPLETE THIS LEVEL AT ALL.
So why not find a new game? Rinse, repeat.
The problem seems two-fold. You might have unrealistic ideas about what startups do. Second, founders fight through shit like this because their motivation is bigger than just getting the "level rush". What's your motivation?
There are all sorts of ways to "hack" yourself here.
1) VERY publicly declare that this project isn't a project-- it's your life.
2) Commit to your friends that you'll run a specific number of experiments (the ones necessary to prove out the model as worth pursuing).
Another thing to consider is to just embrace it. How many have you punted? My dad played bongos in a jazz band, drove a cab, sold veterinary drugs, did electrical work, etc., until he was about 35. He then went on to build companies and ended up in executive management for the last 10 years of his career. You may have to kiss a lot of toads to find your prince... And it MIGHT not be the right decision to force yourself into long toad makeout sessions if it doesn't feel right (okay, that's my best analogy ever).