At work, I have been trying to make synthetic data to test out a technique for making the precipitation data behave a little better. (The problem with precip is that the intensity is skewed, so you need to rescale it, but it doesn't rain every day, which messes up the usual rescaling techniques, so you have to do fancy things, and fancy things are harder.)
The reason why you use synthetic data to test your methods is that then you know what the right answer is, and therefore you can tell whether or not it's actually working. But that means you have to figure out what your method should be doing before you test it to see what it is doing. Which in this case turns out be hard.
Anyway, brains: so yesterday I'm struggling with this, and I was feeling like my mind was all spinning around in little loops, so I ended up going down to the tree plaza outside the building and walked around in small circles talking out loud to myself about the problem for like half an hour.
And it helped. A lot.
It's like, why would that help you think your way through a bunch of abstract mathematical transformations? Wandering and muttering? Really?
But it did. Because brains are weird.