The most challenging aspect of programming, for me, is the ability to turn something over in my head to look at it mentally. To be honest, I can only do that fo

The most challenging aspect of programming, for me, is the ability to turn something over in my head to look at it mentally. To be honest, I can only do that for any length of time on a good day. Writing stuff down is laggy and sometimes lossy, so I am only at my most productive when I can take a complex problem with a lot of little fiddly bits, ingest it mentally, and keep a holistic model in my head. Puzzles provide a good first-order estimation of this ability. Not solving puzzles, per se, but at least analyzing them. A good puzzle requires you to use a part of your brain as "scratch space" to hold hypotheses which are applied to other parts of the puzzle. Can I solve it this way? What if I tweak this part over here? Does this assumption simplify the path to a solution, or miss out on an important subtlety? I think the thing I tried previously was more correct; can I remember how to get back to it? I'm actually not that great at this. Many of my peers are better. Fortunately, I can compensate in other areas (I'm glad to see writing skills called out in the comments, as I like to think I'm a good communicator). I have developed some coping mechanisms (the ability to identify which details in a problem can be succinctly written down and which subset can be stored in my head -- I refer to my notebook as swap space), but I do see the value of puzzle-type questions for evaluating a candidate's ability to do this sort of thing. Does this mean that if you can't do it, you aren't a programmer? Of course not. I happen to be a kernel developer, so I write a lot of code that communicates with other modules, doesn't have much of a UI, and has a multi-dimensional rather than flat structure. Different types of programming need different sets of skills. Find your niche, and don't fret about how you don't fit into other niches as well as you fit into yours. (I interviewed at Google seven years ago and was rejected for being weak at algorithm design. A few years ago I had an aha! moment where I found a simple, elegant, and easy solution to the problem I completely bombed in the interview. Apparently my subconscious had been chewing away at it for literally years. I think solving a question in four years or so is sub-optimal from Google's perspective so I'm sure they were right to reject me, and my inability to do it faster speaks to the deficiencies I mention above. But in the ensuing years I have become reasonably successful in my current line of work, which I enjoy completely. My failure at Google does not keep me up at night.) TL;DR: Puzzles only measure a single axis, so find which axis your greatness lies on and optimize for that.