Detecting subtle effects of persistence in the stock market dynamics

S

stephencrowley

Do any of you guys follow cutting edge research when creating models? I have been following academic journals for over a year now related to "econophysics", financial engineering, quantitative finance.

There seems to be a large disconnect between academic-speak and trader-speak. How many people out there are able to think in terms of both of these worlds?

Here is a paper I've found which I thought might be interesting.

http://www.unifr.ch/econophysics/PH...hp?year=2005&code=physics/0504158&version=abs

Abstract:
The conventional formal tool to detect effects of the financial persistence is in terms of the Hurst exponent. A typical corresponding result is that its value comes out close to 0.5, as characteristic for geometric Brownian motion, with at most small departures from this value in either direction depending on the market and on the time scales involved. We study the high frequency price changes on the American and on the German stock markets. For both corresponding indices, the Dow Jones and the DAX respectively, the Hurst exponent analysis results in values close to 0.5. However, by decomposing the market dynamics into pairs of steps such that an elementary move up (down) is followed by another move up (down) and explicitly counting the resulting conditional probabilities we find values typically close to 60%. This effect of persistence is particularly visible on the short time scales ranging from 1 up to 3 minutes, decreasing gradually to 50% and even significantly below this value on the larger time scales. We also detect some asymmetry in persistence related to the moves up and down, respectively. This indicates a subtle nature of the financial persistence whose characteristics escape detection within the conventional Hurst exponent formalism.
 
I have seen techno speak get so complex a layman cant read it, yet it says nothing new. But as a total laymen I will break it down.

Remember If someone says the markets are not random (Brownian motion) then they must detect patterns (the hurst exponent) in the data set. (Read chart)


So this guy found that within 1 to 3 minutes a nice move had a chance of leading to another move in the same direction more often than what would be anticpated if the market were Random. But this did not hold for longer term charts. He also mentions down moves and up moves are different.

He said he was clever because he did not use the Hurst Exponent in the normal manner he counted conditional probablities. Which probably means something like instead of using a typically cycle filter he used detrend analysis or vice versa.


A trader might say. Hey If I see a new momentum high on my 3 minute spoo chart (and its not during lunch and not on a blow off move into support or resistance) I will buy the retracement to my favorite moving average.
 
Quote from stephencrowley:

There seems to be a large disconnect between academic-speak and trader-speak. How many people out there are able to think in terms of both of these worlds?
More...

The attitude between day traders and economists is one of mutual antipathy.

It's a real shame.

I think the reason that physicists and applied mathematicians are in high demand from investment banks and hedge funds is that they have no stake in this pissing match. Most daytraders have a poor grasp of quantitative methods and a cowboy mentality to make up for it. Most economists have a deeply ingrained institutional bias favoring the efficient market hypothesis which prevents them from developing any empirical understanding of real world markets.

Martin
 
That was a nice concise analysis and exactly what I got from it. I'm thinking of writing some charting software to predict and extrapolate these short term movements. Does anyone know if any existing sofware that already does such a thing?

I'm also trying to think of a way how to apply this same method to the full orderbook instead of the last execution price.

Quote from jem:

I have seen techno speak get so complex a layman cant read it, yet it says nothing new. But as a total laymen I will break it down.

Remember If someone says the markets are not random (Brownian motion) then they must detect patterns (the hurst exponent) in the data set. (Read chart)


So this guy found that within 1 to 3 minutes a nice move had a chance of leading to another move in the same direction more often than what would be anticpated if the market were Random. But this did not hold for longer term charts. He also mentions down moves and up moves are different.

He said he was clever because he did not use the Hurst Exponent in the normal manner he counted conditional probablities. Which probably means something like instead of using a typically cycle filter he used detrend analysis or vice versa.
More...
 
Quote from stephencrowley:

Do any of you guys follow cutting edge research when creating models? I have been following academic journals for over a year now related to "econophysics", financial engineering, quantitative finance.

There seems to be a large disconnect between academic-speak and trader-speak. How many people out there are able to think in terms of both of these worlds?

Here is a paper I've found which I thought might be interesting.

http://www.unifr.ch/econophysics/PH...hp?year=2005&code=physics/0504158&version=abs

Abstract:
The conventional formal tool to detect effects of the financial persistence is in terms of the Hurst exponent. A typical corresponding result is that its value comes out close to 0.5, as characteristic for geometric Brownian motion, with at most small departures from this value in either direction depending on the market and on the time scales involved. We study the high frequency price changes on the American and on the German stock markets. For both corresponding indices, the Dow Jones and the DAX respectively, the Hurst exponent analysis results in values close to 0.5. However, by decomposing the market dynamics into pairs of steps such that an elementary move up (down) is followed by another move up (down) and explicitly counting the resulting conditional probabilities we find values typically close to 60%. This effect of persistence is particularly visible on the short time scales ranging from 1 up to 3 minutes, decreasing gradually to 50% and even significantly below this value on the larger time scales. We also detect some asymmetry in persistence related to the moves up and down, respectively. This indicates a subtle nature of the financial persistence whose characteristics escape detection within the conventional Hurst exponent formalism.
More...

If it's that complicated then it's got to be wrong.

The things I've found that work are very simple! All of my experience points in the opposite direction, 90% of your time will be wasted, imho.

I wish you good luck though.
Keep us informed, if something shows up...
 
Quote from stephencrowley:

...
There seems to be a large disconnect between academic-speak and trader-speak. How many people out there are able to think in terms of both of these worlds?
...
More...
Quote from Sparohok:

The attitude between day traders and economists is one of mutual antipathy.
...

Martin
More...
No problem here at all. The two live in totally different worlds:

(1) The trader struggles in the market in order to extract money from it. As the market has zero credulity, this is a very tough trade;

(2) The academic jostles for academic advancement and consulting fees by telling people stories about the market and in rare cases, offering hypothesis on how to extract money from it. As the gullable public is high on credulity, this is a smoothly running trade.
 
LOL, that brought back memories of a workshop on financial derivatives that I attended a while back.

One presenter - a nuclear physicist that had moved over to "The Dark Side" presented his latest research. During question time, a puzzled member of the audience enquired as to what was this to be used for. The response was "No use. This is purely an academic excercise". (I need to churn out papers to get more research grants).

At least he is honest.

During our luncheon, I had a chat with the same Gentleman - a rather nice and personable chap I might add - quite a refreshing change considering he has been in academia for so long. I asked him if he had made any money in the markets with anything that he had worked on/discovered whilst conducting research in this area. His answer was : "No, I think most academics lose money in the markets".

:)

That is not to say that there aren't people making money using rocket-science. Good luck finding any of that published. In particular, I refer you to Shaw's interview in one of the Market Wizard's books. Lets just say that Dr Shaw was tighter than a fish's rectum when it came to revealing any information about what worked and what didn't - for him/his-firm at least.
 
I completely agree. Simplicity is highly underrated. occam's Razor: "one should not increase, beyond what is necessary, the number of entities required to explain anything".

I've read the paper and the ideas in it are not all that complex. The guy is basically saying that market direction is slightly less than random in the short term (1-3 minutes) and reverses direction more often than not after 10 minutes.

Interesting stuff anyway.. it would be neat to apply this to the evolution of the entire order book. If I come up with something amazing I'll be sure to keep it to myself. :)

--Stephen

Quote from traderlivermore:

If it's that complicated then it's got to be wrong.

The things I've found that work are very simple! All of my experience points in the opposite direction, 90% of your time will be wasted, imho.

I wish you good luck though.
Keep us informed, if something shows up...
More...
 
Yesterday, at 3:15PM crude oil opens at $52, it had closed $49.55 earlier at 2:30PM.

At the time, I was short, totally convinced that oil would continue south.

To makes things easy, 2 QM contracts change $1,000 per $1 move in QM price.

So, a $2.45 move represents $2,450.

So, here I am shitting my pants, trying to sort out what had happened. My first thought, of course, was on the venue of terrorists blew some pipe, somewhere, or worse...

15 seconds later, take my word, for that's what it seemed... prices come down to $49.50 approx, jittering...

So, here again, I'm thinking, what the hell? Is it some uninformed guys (of the tragedy) bringing the price down, or what? Is it my chance to get out? If not, what the hell is going on?

Can't find any related news...

Cashed my chips at $49.67, broke even...

Later on, oil would continue down to hit a low of $48.80 today.

I think I figured it out, some guy with balls placed a big order at the open, took all the stops to $52 and sold... I counted 232 contracts...

My point is, tweezers won't help where you need a bulldozer...

Just trying to help.
 
You must log in or register to reply here.
Back
Top