None. JBadger's numbers don't work. You have to leave out the digit "9" and use in its place the "6" upside down. It is a trick question. Once you understand the 6 for 9 trick, the answer is trivial. 0 1 2 3 4 5 0 1 2 6 7 8
Bingo. And I agree that the nature of such questions are silly for trading interviews. But often, they are asked of entry level technical positions just to see how the interviewee approaches the problem. One way to leave an impression is simply turn the tables on the interviewer and ask him a very difficult question that seems like someone knowledgeable in the field would understand. Most interviewers usually ask if there are any questions from you, and that is your chance to pounce. I've had many a few interviewees impress me that way=) I've also dazzled many interviewers that way. Ask something like, "so what type of robust statistical methods have you found useful in analyzing data; and what are some practical examples where you've used them." Memorize and understand that concept fully, then, regardless of his response (which will likely be shock), turn around and say, "well, I've ran some simulations on blah blah, and found that median statistics are better in that case because blah blah." Or what do you think of Engle's dynamic conditional correlation? I was reading some papers and blah, blah. In this manner, you are taking control of the interview, and showing assertiveness with the ability to impress them with your knowledge.
Well, I think I figured out the 18-coin one: Divide into 3 piles of 6 coins and put any two of these piles on the balance. If they're equal, then keep the 3rd pile, else keep the heavier pile. You have used the balance once and reduced the problem to 6 coins. Now divide those into 3 piles of 2 coins and do the same thing. You have used the balance twice and reduced the problem to 2 coins. Put them on the balance, using it for the 3rd time, and see which is heavier.
If I were asked brainteasers, I would say: I can predict the market with 70% probability, now do you need me or not???!
its not a matter of getting the right answer; its watching the candidate think about the answer. one of you guys answered my cube date question correctly but was a douche about it so no hire imo.
This is actually a good question for programmers, imo, because it has some similarities to how algorithms work. It is not a math question, but a logic question. Also, it can show that you can break a problem down into sub problems recursively. Although, I would not ask if it can be done in three times. I would ask how many times do you need to weigh to identify the heavy object. Everyone gets four, but can you get three? The rest of the questions are idiotic, imo.
Here is a math question that any one that went to high school can answer: For what primes p, is 3p + 1 = x^2 ? Here is one that requires you to probably have a degree: For what primes p, is 3p - 1 = x^2? Here is one that would [probably] win you a Fields Medal: For what primes p is y^p + p - 1 = x^2 ? Notice that (2, 5, 6) works. 2^5 + 5 - 1 = 6^2. Are there infinite number of them? Or is 5 the only prime that makes the statement true. What about y other than 2? Does y have to be prime? These are math questions (Number Theory). The rest are not math questions. In fact, I would wager that some of the greatest mathematicians alive would not be able to answer the "math" questions asked here. Mathematics is about ideas, not dog and pony circus tricks.
The only question they want answered is- "How much are you going to fund your account with". -------------- Answer- If "x" is > or = "firm minimum" then YOU'RE HIRED.
Here are real mathematicians talking about a gorgeous problem: http://video.google.com/videosearch...Mqzyye0H&sa=X&oi=video_result_group&ct=title#