I think it depends on which version of the Efficient Market Hypothesis you're talking about. This EMH <-> P=NP paper deals specifically with the weak EMH. And not everyone even agrees on what the right definition for weak, semi-strong, and strong version of it are.
I don't have strong opinions on the EMH, realizing that it is a complicated subject and I'm not an economist. However, I tend to lean towards agreeing with this essay, which indicates that some form of the weak EMH probably holds but stronger forms do not:
According to this, it's the semi-strong version that will have to be rejected after the financial crisis (and may have been widely accepted before it). While the weak form still holds up to a lot of empirical data.
So the claim in this P=NP paper is far bolder, and upon reading it I find a lot of the reasoning seems skethcy... it's an interesting analogy, but I just don't think I buy the idea that you have to search the entire possible exponential pattern space in order to identify patterns. I think both people and computers have more efficient ways of pattern recognition, whereas the case he's thinking about is a worst-case scenario which would almost never be relevant in practice. At least that's my take on it; could be wrong.
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
timeasmymeasure
no subject
Date: 2010-03-07 06:41 pm (UTC)I don't have strong opinions on the EMH, realizing that it is a complicated subject and I'm not an economist. However, I tend to lean towards agreeing with this essay, which indicates that some form of the weak EMH probably holds but stronger forms do not:
http://johnquiggin.com/index.php/archives/2009/01/02/refuted-economic-doctrines-1-the-efficient-markets-hypothesis/
According to this, it's the semi-strong version that will have to be rejected after the financial crisis (and may have been widely accepted before it). While the weak form still holds up to a lot of empirical data.
So the claim in this P=NP paper is far bolder, and upon reading it I find a lot of the reasoning seems skethcy... it's an interesting analogy, but I just don't think I buy the idea that you have to search the entire possible exponential pattern space in order to identify patterns. I think both people and computers have more efficient ways of pattern recognition, whereas the case he's thinking about is a worst-case scenario which would almost never be relevant in practice. At least that's my take on it; could be wrong.