Brain teasers in a job interview generally show that the interviewer has
a) no clue about his technical/operational subject. If an interviewer is not able to judge whether a person is qualified for the job by having a good meeting, listening to his/her previous experience and the things he/she has done already in real live, it is already a sign that something went wrong in bringing this person on board. I know that “brain teasers” are an increasingly found bad habit and I presume that this is coming from this idiotic business consultant type of persons (a la assessment centers). Which again would support the thesis that they do not have real experience in their jobs and thus need such strange methods like brain teasers and puzzles – if the interviewer lacks the technical/operational qualification there the only method of judgment obviously is a “brain teaser”.
b) significant deficits in management skills: I want people who are honest, well trained and diligent … and the last method to find that out is a “brain teaser”. It is even questionable to find out whether a person is “intelligent” (what is the definition for intelligent ?) by such things. I don’t pay people solving puzzles and playing game (my little son is doing that, and he is in the right age for it)
But even more important in my opinion – it shows a considerable amount of disrespect concerning the interviewee. Im my opinion it is extremely unprofessional to play the “who has the longest …” game at work.
Fortunately I do not have to go these types of interviews (and hopefully I will not have to) … but if so, before working for somebody who runs around throwing “brain teaser” questions at people (not for fun but with the intent of making “judgment”) I would rather change my career path.
Do you ask a Fund Manager if he is good at “brain teasers” ? Probably not – you look whether the guy is successful in his track (which also involves alot of social intelligence to convince people) and whether his concepts are sound.
So when would there be the need to ask him a brain teaser question : only if you did not understand his concept and you were not able to make judgment from his track record, right?
Maybe … the questions which I should ask myself as a manager is, whether I want people who “accommodate” to every silly idea I have or whether I want to have a team of experts whom I can rely upon also when I am not in the office (which implies that the manager is capable leading a highly qualified expert group) … it is a question of leadership.
Clearly, a team needs a leader and the leader has the final decision … however it should be considered as a qualification of the leader to differentiate between things that are important (and thus his decisions are mandatory to all team members) and things that are just “baloney” (like “brain teasers”) …
I think, that if you are a quant expert you will always get a job … and those companies which do not consider you (despite your expertise, work experience and innovation capacity) just because you restrict your own thinking to that which is appreciated by your boss, you probably do not want to work for …
The assumption that the interviewer IS good is an idealized assumption (a bit similar to a model optimized on past data 
. If you look at overall fund tracks (cleansed of survivorship bias) the assumption that the interviewers ARE has statistically not too much evidence, does it !? (you could say the sales guys ARE per definition good …). That brain teasers are common maybe, which does not mean that they are bringing wrong results. The thing I m afraid of is, that the risk of a false negative, i.e. sending a real crack with a huge amount of knowledge away, just because he had a bad day or another one has heard the question before already is too high. Nobody knows what the real ratios in terms of false positive or false negatives with these types questions are … however they are used without actual figures at hand – and I suspect that many “interviewers” work like that – using techniques with no sufficient evidence of success, hit rates, and potential blow-up cases – also in their financial software. Indeed thats the way of creating a couple of mil loss easily …