I'm going to take the contrarian point of view here: partially for the sport of it, and partially because I genuinely think there's value in puzzles and brainteasers -- though I do think the value is limited, circumstantial, and merely one datum out of many in an interview.
These questions are about having a logical process, regardless of where it gets you, and spelling that process out in a sequential fashion. It's more important to communicate "I do X, and check for Y. Then I do Y, and check for Z" than it is to arrive at a reasonably accurate Z.
The added component of pressure (i.e., you've only got a few minutes, in the uncomfortable setting of an interview, in which to answer the question) helps (or allegedly helps) detect your ability to remain calm and logical under tight deadlines.
Basically, the question is asking "Under pressure, and faced with a seemingly insurmountable problem, will this candidate give up, or will he try to address obstacles in a rational, collected, and systemized fashion?" They're essentially variants on the Kobayashi-Maru test of Star Trek fame.
Furthermore, I don't agree with the apparent consensus that oddball "puzzle" questions would be better replaced with more technically relevant brainteasers. That's not the point. You have technical questions to test your technical literacy. These questions are trying to test your character. (Brainteasers are often non-technical precisely because the interviewer doesn't want you to be able to fall back on existing knowledge as an escape hatch from the question). Some brainteaser/puzzle questions are much more inane than others, but the exercise itself isn't entirely worthless.
I have never asked a puzzle question in an interview. If I were compelled to do so, I would never hire someone purely on the basis of his or her performance on one. It's one of the least important variables in a hiring process. But I'd consider with some skepticism anyone who outright refuses to answer a brainteaser, or who gets totally flummoxed by one, or who gives up without at least attempting to work out an approach.
Ultimately, no single type of interview question is flawless. That's why you have a wide variety in your arsenal. You've got technical questions, case questions, puzzle questions, whiteboard questions, and even the oft-derided "Tell me about a time when..." questions. In isolation, none of these types is sufficient. In combination, they test different aspects of a candidate's thought process and preparation.
Brainteasers and problems are fun and I love doing them. I just think it's not appropriate for an interview, especially when you catch an interviewer presenting you with questions that are simply taken from known 'problem banks'. I could have easily pretended to work through them like I've never seen them before but I think it's a lot more truthful to call them out and say 'I know all your questions'.
"especially when you catch an interviewer presenting you with questions that are simply taken from known 'problem banks'"
Sure, but then, almost all interview questions can be similarly gamed. If an interviewer asks me to "walk him through my resume," or to "Tell him about a time when I did X," does he really expect that I haven't thoroughly prepared and rehearsed those answers? And if I truly haven't, then that's saying quite a bit about how seriously I take the process.
I'm not suggesting that brainteasers are as valid as other question types. There are plenty of question types that better get to the heart of the candidate's aptitude, experience, skills and capabilities, domain knowledge, and so forth. But everything can be rehearsed for. I don't think that's a valid categorical critique of the puzzle type.
Better arguments against puzzle questions, IMO, are: 1) They might select for good bullshitters, as opposed to good thinkers; 2) The opportunity cost of spending time on a puzzle question is the time that could be spent on a more relevant question; 3) Some people just aren't good at (or interested in) puzzles, but that doesn't mean they're not awesome on the job -- ergo, passing on a candidate because he flunked a puzzle is much sillier than passing on a candidate because he flunked a technical exercise. (Conversely, loving a candidate because he kicked a puzzle's ass is even sillier).
Good points. That is probably why people are often not successful on their first interviews. It takes a few to get into "interview mode" where you have good answers for common questions about your CV.
I was quite dismayed when I called them out and they didn't come back with new questions or made up some on the spot. They simply decided to drop that part of the interview and move to the next section.
I've seen too many good people (friends and students of mine) miss out on jobs simply because they failed these type of problems.
"I've seen too many good people (friends and students of mine) miss out on jobs simply because they failed these type of problems."
With this I would absolutely concur. I have more than a few friends who are smart, and who are rock-solid performers, but who just don't have a knack for brainteasers. I would hope than an interviewer is able to suss this out in an interview, i.e., make the determination that a person is great in every aspect, would probably be a great fit, but just happens to suck at puzzle questions. Sadly, I'm sure it's more often the rule than the exception that an interviewer will reflexively pass on someone who flubs a puzzle.
And I agree with you that puzzle questions require flexibility not only on the part of the interviewee, but also on the part of the interviewer. If you're just tossing out questions based on a book of canned puzzles, don't be surprised when you get canned responses. The exercise becomes entirely meaningless when that happens, and there's really no use in proceeding with it if it does.
These questions are about having a logical process, regardless of where it gets you, and spelling that process out in a sequential fashion. It's more important to communicate "I do X, and check for Y. Then I do Y, and check for Z" than it is to arrive at a reasonably accurate Z. The added component of pressure (i.e., you've only got a few minutes, in the uncomfortable setting of an interview, in which to answer the question) helps (or allegedly helps) detect your ability to remain calm and logical under tight deadlines.
Basically, the question is asking "Under pressure, and faced with a seemingly insurmountable problem, will this candidate give up, or will he try to address obstacles in a rational, collected, and systemized fashion?" They're essentially variants on the Kobayashi-Maru test of Star Trek fame.
Furthermore, I don't agree with the apparent consensus that oddball "puzzle" questions would be better replaced with more technically relevant brainteasers. That's not the point. You have technical questions to test your technical literacy. These questions are trying to test your character. (Brainteasers are often non-technical precisely because the interviewer doesn't want you to be able to fall back on existing knowledge as an escape hatch from the question). Some brainteaser/puzzle questions are much more inane than others, but the exercise itself isn't entirely worthless.
I have never asked a puzzle question in an interview. If I were compelled to do so, I would never hire someone purely on the basis of his or her performance on one. It's one of the least important variables in a hiring process. But I'd consider with some skepticism anyone who outright refuses to answer a brainteaser, or who gets totally flummoxed by one, or who gives up without at least attempting to work out an approach.
Ultimately, no single type of interview question is flawless. That's why you have a wide variety in your arsenal. You've got technical questions, case questions, puzzle questions, whiteboard questions, and even the oft-derided "Tell me about a time when..." questions. In isolation, none of these types is sufficient. In combination, they test different aspects of a candidate's thought process and preparation.