Forums  > General  > Markets Are Efficient If P=NP  
     
Page 1 of 1
Display using:  

lmog


Total Posts: 134
Joined: Mar 2010
 
Posted: 2018-06-04 23:40
How long does it take before your crank alarm goes off?



https://arxiv.org/abs/1002.2284

Check out the video of the author: https://www.youtube.com/watch?v=7iOJZZFDKpc

katastrofa


Total Posts: 451
Joined: Jul 2008
 
Posted: 2018-06-04 23:45
I think I've seen this paper when it came out. Why do you think it's a crank?

lmog


Total Posts: 134
Joined: Mar 2010
 
Posted: 2018-06-04 23:59
Watch the video...does it come across as compelling even without a detailed analysis of the claims?

pj


Total Posts: 3400
Joined: Jun 2004
 
Posted: 2018-06-05 10:39
Well, Black Scholes model is efficient.
Therefore .

On the other hand the fractional brownian motion models are not.
Therefore

Confused

The older I grow, the more I distrust the familiar doctrine that age brings wisdom Henry L. Mencken

ronin


Total Posts: 317
Joined: May 2006
 
Posted: 2018-06-05 15:08

It's actually a cute paper. I like it - thanks for putting it up.

"There is a SIX am?" -- Arthur

lmog


Total Posts: 134
Joined: Mar 2010
 
Posted: 2018-06-05 17:38
Doesn't trigger just a little skepticism?

katastrofa


Total Posts: 451
Joined: Jul 2008
 
Posted: 2018-06-05 20:24
WTF would I judge the paper based on how the guy looks on Youtube?

lmog


Total Posts: 134
Joined: Mar 2010
 
Posted: 2018-06-05 21:26
It isn't how he looks, it is about what he saying and how he is saying it. Anyone critical uses that to judge credibility.

ronin


Total Posts: 317
Joined: May 2006
 
Posted: 2018-06-06 15:08

He looks like an academic. He talks like an academic too.

What's the problem @lmoq?


"There is a SIX am?" -- Arthur

katastrofa


Total Posts: 451
Joined: Jul 2008
 
Posted: 2018-06-06 18:48
@lmoq

Feel free to cite what he's saying in the paper and refute it.

lmog


Total Posts: 134
Joined: Mar 2010
 
Posted: 2018-06-06 20:32
Many times it isn't worth getting into the claims of the paper, it really seems to be the case here. Academics are typically are a lot more concise about subjects that they know about.

Look at this comment in the context of a infamous "proof" of P != NP. The paper hits Scott Aaronson's ‘Ten signs a claimed mathematical breakthrough is wrong’ (a crank checklist much like John Baez's). The exposition on basic concepts is belabored, the conclusion is deep, the methods used don't seem to be nearly powerful enough to get to that deep conclusion.

I've actually read through the paper and can see all the stuff that doesn't make sense, I'm surprised that this paper hasn't been called out more clearly.

katastrofa


Total Posts: 451
Joined: Jul 2008
 
Posted: 2018-06-06 22:06
You mean the comment which calls the preprint "enjoyable", "mathematically orthodox" and says that "it embraces conventional nomenclature, asserts well-posed definitions, employs viable proof strategies, and claims theorems that do not contradict known results"?

Are you that dense that you don't understand that the comment you linked to is an *endorsement* of the article you're trying to piss on?

lmog


Total Posts: 134
Joined: Mar 2010
 
Posted: 2018-06-06 23:32
I don't think you are reading the comment critically (admittedly, Sidles' writing can be hard to understand):

Yet Maymin’s preprint does (very enjoyably) flaut at least the following five of Scott’s ten rules: Rule 1: the manuscript isn’t TeX’d, Rule 3: the approach seems to yield something much stronger and maybe even false (namely, that implication the postulates of free-market economics are wrong), Rule 5: the approach conflicts with a known result (namely, the empirical evidence that markets often are efficient), Rule 8: the manuscript wastes lots of space on standard material (the reductions are trivial, the exposition belabored), Rule 10: the techniques just seem too wimpy for the problem at hand (undergraduate-level complexity theory suffices).

It’s not easy to write a short essay about why Maymin’s preprint is receiving attention despite transgressing these rules.

katastrofa


Total Posts: 451
Joined: Jul 2008
 
Posted: 2018-06-06 23:55
"Yet Maymin’s preprint does (very enjoyably) flaut at least the following five of Scott’s ten rules: Rule 1: the manuscript isn’t TeX’d, Rule 3: the approach seems to yield something much stronger and maybe even false (namely, that implication the postulates of free-market economics are wrong), Rule 5: the approach conflicts with a known result (namely, the empirical evidence that markets often are efficient), Rule 8: the manuscript wastes lots of space on standard material (the reductions are trivial, the exposition belabored), Rule 10: the techniques just seem too wimpy for the problem at hand (undergraduate-level complexity theory suffices)."

Let's say he flouts all these rules (though you'd have to offer some arguments about #10). So what? Why should I care that the manuscript isn't in TeX? What kind of snobbery is that?

Anyway, back to the paper. The abstract says something which seems to have gotten your pants (and possibly Sidles') in a twist: "Specifically, markets become increasingly inefficient as the time series lengthens or becomes more frequent." This is not a controversial opinion, IMHO. It is consistent with physical intuition (processing and transmitting information takes non-zero time).

Are you able to point out an actual error in the paper? Discussions about style or arbitrary rules made up by some bloggers are boring.

lmog


Total Posts: 134
Joined: Mar 2010
 
Posted: 2018-06-07 00:16
Ah I see why you aren't seeing it now. You think Scott Aaronson is some blogger.

katastrofa


Total Posts: 451
Joined: Jul 2008
 
Posted: 2018-06-07 00:25
I know who Scott Aaronson is ;-)

lmog


Total Posts: 134
Joined: Mar 2010
 
Posted: 2018-06-07 00:38
Alright so you've gone through the paper then and think it is made of sound reasoning?

katastrofa


Total Posts: 451
Joined: Jul 2008
 
Posted: 2018-06-07 00:47
Yes. Can you prove otherwise?

pj


Total Posts: 3400
Joined: Jun 2004
 
Posted: 2018-06-07 06:41
Ahem. What about my observation in
my previous post?

His reasoning is somewhat similar to the
old quip.
Consciousness is complicated and mysterious.
Quantum mechanics is complicated and mysterious.
Therefore consciousness is quantum mechanical.

I did enjoyed his article though. Just I don’t buy it.

The older I grow, the more I distrust the familiar doctrine that age brings wisdom Henry L. Mencken

katastrofa


Total Posts: 451
Joined: Jul 2008
 
Posted: 2018-06-07 12:32
Whose old quip?

gaj


Total Posts: 23
Joined: Apr 2018
 
Posted: 2018-06-07 13:46
His definition of market efficiency is roughly: given a price series of length n there is no polynomial time algorithm to find statistically significant profitable strategies. This definition has nothing to do with markets or trading. It's purely a mathematical statement about of sequences of numbers. The paper looks logically sound (although I don't understand section 2 at all), but the connection to market efficiency is a bit of a stretch.

pj


Total Posts: 3400
Joined: Jun 2004
 
Posted: 2018-06-07 13:59
> Whose old quip?
Leonard Mlodinow's in discussion with Deepak Chopra.
You can find it on Youtube.

+1 gaj post

The older I grow, the more I distrust the familiar doctrine that age brings wisdom Henry L. Mencken

lmog


Total Posts: 134
Joined: Mar 2010
 
Posted: 2018-06-07 18:01
I've been trying to think of the most succinct holes in the paper, given it is filled with little things that don't make sense.

Given gaj's proper summary of Maymin's paper "there is no polynomial time algorithm to find statistically significant profitable strategies", it could still be that there are methods that are "quick" (as Maymin uses the term) that have super fast average case complexity.

Maymin brings up SAT solvers (which can solve many practical and huge instances of an NP problem very very quickly), doesn't mention DPLL in his exposition of how they work and doesn't explain how this doesn't defeat his argument.

Of course, why are we even here when the connection to "market efficiency" doesn't seem clear, but, there it is in the title.

lmog


Total Posts: 134
Joined: Mar 2010
 
Posted: 2018-06-07 20:58
pj's quote is also right on and is exactly what "smells bad" about this paper.

x is complicated, y is complicated....by completely superficial analogy we might ask "is x related to y?". Asking is ok. Many good ideas, presumably, begin with superficial analogies. The analogy doesn't seem to be deep here. Certainly not as much as the title suggests.
Previous Thread :: Next Thread 
Page 1 of 1