Do I really support the Kobayashi Maru in interview questions?

by Michael S. Kaplan, published on 2005/04/23 04:30 -04:00, original URI:

When I wrote about an interview question that I would ask (after going on about one that I would not ask), I got some interesting feedback.

One of the more interesting questions I got via e-mail was the subject line of this post. Jen (the developer asking) had managed to find a question that was the perfect balance of insight and geeky Star Trek references, and I knew I would have to post about it (even if she didn't put it in a comment; I am convinced that some people simply do not do that).

Anyway, this developer was looking at the concept of asking a question that you honestly expect the candidate to not be able to answer in the time alloted, and whether looking at the way they approach the problem. To her it felt like the sort of "no win" scenario that the whole Star Trek "Kobayashi Maru" test was all about.

For those not as geeky as her or I, it is described well in a Wikipedia article:

The Kobayashi Maru is a third-class neutronic fuel carrier-ship in the fictional Star Trek universe which serves as the basis of training exercise in which command division cadets were presented with a "no-win" scenario as a test of character at Starfleet Academy. The name is Japanese, and loosely translates as the ship named Kobayashi, with Kobayashi (小林) meaning small forest and being a common family name. Maru (丸) simply is a suffix for Japanese ship names.

The theory behind the test, in the Star Trek universe, is that a Starship Captain would really only ever get to face such a situation once in their lives, and so it was basically just a test of character to see how they do in that situation. As the article goes on to mention:

James T. Kirk took the test three times while he was at the academy. Prior to his third attempt, Kirk surreptitiously reprogrammed the simulation computer to recognize his name and have the Klingons react with fearful admiration. He justified it by arguing that putting cadets in a no-win situation was cheating, and so he had to cheat in return. He received a commendation for original thinking.

Jen was really somewhat bothered that I was doing this sort of thing to an interview candidate -- giving them a "no-win" scenario and then watching them squirm and try to solve the problem, as a test of character. In the end, she reasoned, I was not interviewing for Starfleet Academy; I was interviewing for Microsoft. Was I aware if this fact?

I decided to answer Jen's question in a post since I am sure there are others who would appreciate the geekiness of the question and maybe of the answer. :-)

As I stated there, I seldom do the "top-down" sort of question because I am bothered by the very same things that bothered her (though I had never made the 'Kobayashi Maru' connection before she asked it!). Putting the candidate in a question they could not fully solve can be incredibly stressful on a day in which a candidate is already incredibly stressed. The fact that I did well when Dave asked me such a question (or that my colleague Eric did well the one time I asked such a question) is really irrelevant -- in both cases we were surprised that a "James T. Kirk" had been pulled and someone managed to beat the "no win scenario". In a way I was cheating when I beat it, so I was indeed like a "James T. Kirk" who did not even need to take the test three times, while Eric just genuinely answered the question well, so he was a hire in my book1 (and he was hired, so he must have impressed other people too!).

Truthfully, as I stated in my post, I usually prefer a "bottom-up" type question like the second one I gave.

The reason is that with such a question, the candidate gets to answer as many questions as there is time for them to answer, and at each step of the way they get to know that they are being rewarded by trying to make the question a bit harder. They never get to "finish" the question in the sense of totality but they never get the feeling that they were supposed to (as I never really frame the problem in a way that makes them think they have to).

Much less pressure, they get to feel like they are answering questions, and if they do very well and go far in the process, I tell them that sincerely so they know they are doing quite well compared to other candidates in the same situation (the inverse is not true; if they are not doing well I do not tell them that they suck compared to others!).

Interviews at Microsoft are not really about presenting a candidate with a Kobayashi Maru, a no-win scenario, and seeing what they do with the scenario. They are about finding the smart people who we believe will do well both for and at Microsoft.

Jen, I hope that answers your question. And that you were truly not discouraged from interviewing on the basis of the post (I was hoping it would do exactly the opposite!). You should also look at other even cooler blogs about MS and interviewing, like the ones from Heather and Gretchen, as they have a lot more information and are a lot less likely to give the wrong impression!


1 - And not just because of the fact that he beat a "no-win scenario" but because of the skill and insight he bought to the various facets of the question caused me to realize that this was a mind that it would be a good idea to have working for us!


This post brought to you by "ש" (U+05e9, a.k.a. HEBREW LETTER SHIN)
Because this post is קשר לפסח in anticipation of the festivities that start less than 24 hours from now)

# Jason Coyne on 23 Apr 2005 10:51 AM:

When I interviewed at Microsoft, I had a HORRIBLE question asked to me. Rather than being something that nobody could solve, so figure out your thinking process (like, how many gas stations in LA), mine was an actual solvable problem, but I was not given enough information to solve the problem :

If a boat is in a lake, and a fish is caught, when the fish is pulled into the boat, does the level of the lake rise, or fall.

You need to know the boyancy (and shape) of the boat, and the volume and mass of the fish.

Without these peices of information, the question is unsolvable, because both senarios are possible.

I think the value of these types of questions has really decreased over time, particularly since they aren't "suprise" questions anymore. Everyone knows they are coming.

I did get a much better technical question from a different interviewer : how do you determine if a linked list is circular, using the minimum amount of memory possible (time to execute not a factor)

go to newer or older post, or back to index or month or day