The rule-of-thumb of factoring out only when you've written the same code three times rarely gets a chance here, because as soon as you notice a regularity, and you think critically about it, your next experiment breaks that regularity.
It's tempting to create reusable modules, but for one-off exploratory code, for testing hypotheses, it's far more efficient to just write it.
I have tons of examples of code where I did the simplest thing to solve the problem. Then later needed a change. I could refactor the entire thing to add this change or just hack in the change. Refactoring the entire thing takes more work than the hack so hack it is unless I forsee this is going to matter later. Usually it doesn't
That’s just anecdote, just like mine. Even simple lack of experience or lack of skills can cause that (which were definitely in my case). Also, I’m quite sure that a terrific coder can create maintainable code faster than an average one bad code. That’s why I asked some statistical data about that.
It's tempting to create reusable modules, but for one-off exploratory code, for testing hypotheses, it's far more efficient to just write it.