Citadel/Two Sigma/DE Shaw/Tower/Jump/HRT Quant Intern

arjiang 发表于 2019-11-3 13:23
恭喜楼主. 请问你在HRT的2轮电面时有在线写码么, 还是仅口头回答算法.  on-site时有写c++么? 同在准备HRT面 ...

我面的是他们家今年新开的 Mid-frequency Algo intern, 所以面试侧重点也与他们家传统的高频交易很不同. 电面时没有在线写码. onsite 不好意思不能说更多了.

andrew_shen 发表于 2019-11-3 14:02
我面的是他们家今年新开的 Mid-frequency Algo intern, 所以面试侧重点也与他们家传统的高频交易很不同.  ...

多谢楼主! 祝实习科研两顺利

学长这ID太显眼了hhhh膜

P(X>3Y|X>0). X,Y are i.i.d. standard normal.
老哥请问这种题网上有没有什么讲解？谢谢！

Mathblessing 发表于 2019-11-4 08:20
学长这ID太显眼了hhhh膜

😉

kair 发表于 2019-11-4 09:05
P(X>3Y|X>0). X,Y are i.i.d. standard normal.
老哥请问这种题网上有没有什么讲解？谢谢！

I think the key is to first show that given W = aX + bY, cov(W, X) = 0.
Then the problem becomes: P(F(X,W)|X>0) = P(F(X,W), X>0) / P(X>0).
P(F(X,W), X>0) given that these are not correlated, you may be able to find it by integration (?).

But I doubt you are given enough information since to get cov(W, X)=0 you will first need to have cov(X, Y).

andrew_shen 发表于 2019-11-3 20:27
😉

感觉猜到是谁了。。。那这么强就make sense了，祝将来越来越牛

philipBKXN 发表于 2019-11-5 00:34
I think the key is to first show that given W = aX + bY, cov(W, X) = 0.
Then the problem becomes: ...

X, Y are i.i.d. standard normal means cov(X,Y)=0.
The problem is actually pretty simple: P(X>3Y|X>0)=P(X-3Y>0,X>0)/P(X>0) by definition. The denominator is 1/2 obviously. To compute the numerator, draw the intersection of the two half-planes represented by X-3Y>0 and X>0, then use the rotational symmetry of standard bivariate Gaussian. The final result will be proportional to the angle spanned by the two half-planes intersection.

太厉害了，我膜拜一下

请问一下楼主 DE Shaw的technical部分有考察leetcode嘛？ 还是主要就是概率统计方面的问题呢？