|
|
发表于 2009-7-26 09:08 PM
|
显示全部楼层
本帖最后由 seabiscuit 于 2009-7-26 22:12 编辑
bobcat 是对的，尽管对多数散户来讲不practical。所以DJ的计算是实用的。
jsl 发表于 2009-7-26 20:47 
定义无所谓对错， 但得先说清楚想定义的东西是什么， 如果大家讨论的东东都不一样， 何谓对错？ I think for one single trade, Duoji's calculation makes much more sense. Consider an extreme case (and of course let's take it simple). Say a stock is now $1. You sell short it for $1. After some time, it becomes $1M per share (What a cursed short sell!!!).
Based on Duoji's calculation, the return is (1-1,000,000)/1, which is a very large negative number. This reflects the reality that you have a really large loss.
Based on the other way, the return is (1-1,000,000)/1,000,000, approximately -1, which is very close to 0, though negative. This is very anti-intuitive.
That's why Duoji's calculation is better, in my opinion. |
|
狗仔卡
发表于 2009-7-26 06:27 PM
. That is why it is called "statistical" arbitrage, as opposed to "deterministic arbitrage". Gosh, I thought statistical arbitrage was to use statistical methods for arbitrage. What a misleading name! 
