Algorithms Prove No Match For Market Tumble

from the all-in-the-data dept

Thu, Aug 9th 2007 8:11pm — Joseph Weisenthal

Earlier this year, famed inventor (among other things) Ray Kurzweil announced that he was getting set to launch a new investment company that would employ advanced mathematical techniques to discern patterns in the stock market. Of course, just because Kurzweil is widely considered to be a genius, it doesn't mean he'd have any midas touch when it comes to the stock market. Furthermore, quant funds, as they're known, aren't novel. Lately, in fact, they've been getting slaughtered as the market swings about in unpredictable ways. Top tier investment bank Goldman Sachs has announced that its closing one of its premier quant funds in light of recent losses. Obviously computers have become an indispensable tools in modern day investing, but no algorithm or mathematical genius can guarantee good performance in all markets.

Thank you for reading this Techdirt post. With so many things competing for everyone’s attention these days, we really appreciate you giving us your time. We work hard every day to put quality content out there for our community.

Techdirt is one of the few remaining truly independent media outlets. We do not have a giant corporation behind us, and we rely heavily on our community to support us, in an age when advertisers are increasingly uninterested in sponsoring small, independent sites — especially a site like ours that is unwilling to pull punches in its reporting and analysis.

While other websites have resorted to paywalls, registration requirements, and increasingly annoying/intrusive advertising, we have always kept Techdirt open and available to anyone. But in order to continue doing so, we need your support. We offer a variety of ways for our readers to support us, from direct donations to special subscriptions and cool merchandise — and every little bit helps. Thank you.

–The Techdirt Team

Filed Under: algorithm, quant

14 Comments

If you liked this post, you may also be interested in...

Reader Comments

Subscribe: RSS

View by: Time | Thread

John Doe, IT Administrator, 9 Aug 2007 @ 11:54pm

Algorithms, what's that?
[ link to this | view in thread ]
Anonymous Coward, 10 Aug 2007 @ 2:24am

Mathematical Geniuses are rubbish investors
Black and Scholes invented option pricing. They continue to lose money when they invest following their models (e.g. LTCM).

On the other hand, just because one algorithm is bad, it doesn't mean all algorithms are bad.
[ link to this | view in thread ]
Doug, 10 Aug 2007 @ 5:04am

Black and BS ..
Black and Scholes, more like Black arts, People who bet other peoples money on the stock market by pretending to have a method of predicting the future, through the use of quasi-mathematical terms borrowed from calculus (1666), the number of the beast !!!

Of course if the market only used the same data as the quant funds then the predictions would be .. predictable. What they should do is outlaw people reading newspapers or web articles, like this one.

In case anyone don't realise it by now, noone can predict the future not even our current version of the priesthood in white, scientists. You can't predict the future from past events as the future is unpredictable.

A financial mag here one did a test, they set up a fund manager and amonkey with a pin to 'invest' in a protfolio for a set period. The monkey won !!!

Might as well wave a dead chicken at it. ( Slaughter me a calf and examine the entrails therein that I may place a put option .. :)
[ link to this | view in thread ]
knome, 10 Aug 2007 @ 5:44am

http://en.wikipedia.org/wiki/Pi_(film)
[ link to this | view in thread ]
OKVol, 10 Aug 2007 @ 6:45am

This just proves a point
There is no logic to the stock market. It is driven by emotion, gut feelings, and hunches.

It still is true that a great company can be forced to close because the stock market thinks they aren't good.

Maybe Chrysler has a chance now that they are privately held. They can look past the short term quarterly based goals of a public company.
[ link to this | view in thread ]
Anonymous Coward, 10 Aug 2007 @ 7:21am

The problem is in the details
One possible reason these mathematical models fail in actual markets is in the assumptions made. One of the fundamental assumptions in the Black-Scholes market model is that stocks are normally distributed (that stocks stay, on average, close to a mean price). Anyone who knows anything about stock markets and trading knows this a pretty strong assumption (and could be very wrong), which could explain why so many "impossible" disasters occur in finance markets.

While I agree that reading markets successfully can take some luck and gut feelings, cold hard algorithms could also do some good, as long as they assume the "right" things about markets. The problem is this: good math models necessarily include more complex assumptions (and so are tougher to analyze) to be more realistic. The key to a good algorithm is finding a trade-off in being realistic and being able to actually do something useful.
[ link to this | view in thread ]
Matt, 10 Aug 2007 @ 7:52am

It's a paradox (and so are my shoes)
The most unknowable factor is how the application of a model can fundamentally change the system it is modeling when that system is aware of itself being modeled.
Philosophically speaking, could the stock market benefit from such a model? Could such a model even exist, given that part of its existence is defined by what it defines itself? If Kurzweil is successful, will existence cease to be? Or, if his model is never used will the people who would have sworn by it instead be swearing by tea leaves as a stock market predictor?
Matt
[ link to this | view in thread ]
Anonymous Coward, 10 Aug 2007 @ 8:20am

Some would question your claim that Kurzweil is a genius. Some would prefer to classify him as a charlatan. As far as I know, he does not personally do any technical work (at least not any more, if he ever did) and publishes no technical documents. At best he outlines an objective and hires people to actually implement his vague ideas. In at least one of his books (the one I have read) he makes several egregious technical errors; obvious errors to anyone with even rudimentary understanding of the subjects. He feels to have the authority to talk on many subject clearly outside his expertise (whatever that is). He has a health book who's promotional material suggest that it would be better sold for "4 easy payments of $14.99" than at Barnes and Noble. It could be said that he is a mountebank, making money from the attention received by making wild claims.

Even his futurist predictions are highly criticized. In any case, these predictions are speculation and not scientific work, and predictions not withstanding his own technical merit, particularly in the field of AI, appears severely limited.
[ link to this | view in thread ]
Anonymous Coward, 10 Aug 2007 @ 8:39am

Re:
I'm with the other cowardly critic of Kurzweil. I saw him debate Gelernter at MIT about the coming "Rise of the machines", and I have never seen a more ridiculous and irrelevant spectacle.
[ link to this | view in thread ]
Lorenzo, 10 Aug 2007 @ 10:40am

You can't predict the future...
...you can only guess. If you guess the market will go up, you will likely be right. Why? Because the market bias is up. You can buy a given stock or better yet, an index, and at some point in the future, the index will have a greater value. That's why a chimp can toss darts at a board of stock listings and make money. Pick the stocks a chiken poops on, you'll make money. The illusion of predicting the markets direction gets exposed when the short term bias becomes downward. No one, absolutely no one can predict the future, but you can pretend to do so, by riding the markets bias upward. And better yet, if you have the right swagger and lingo, people will call you a genius.
[ link to this | view in thread ]
Anonymous Coward, 10 Aug 2007 @ 11:18am

Re:
my guess is funny quotes from Al Gore, you know, like Qualisms
[ link to this | view in thread ]
Anonymous Coward, 10 Aug 2007 @ 11:24am

Re: You can't predict the future...
...without psychohistory
[ link to this | view in thread ]
Joe Smith, 10 Aug 2007 @ 12:28pm

Re: The problem is in the details
One possible reason these mathematical models fail in actual markets is in the assumptions made. One of the fundamental assumptions in the Black-Scholes market model is that stocks are normally distributed

I understand one of the problems LTCM got into was that the model assumed that the market would stay liquid at all times. This was the same assumption the portfolio insurance schemes made in 1987 - with the same result when the assumption turned out to be untrue.
[ link to this | view in thread ]
StrataGenie, 10 Aug 2007 @ 4:04pm

Algorithmic Trading
There are some institutions in the algorithmic trading business that turn a substantial profit every year - for nearly the past twenty years. (for example, http://en.wikipedia.org/wiki/Renaissance_technologies) A good algorithm senses market direction continuously, moves into and out of positions faster than you can hit a single key, and profits directly from the fear and greed that make investing notoriously difficult for humans.

StrataGenie is a new system which draws together the combined power of multiple computers over the internet to form a grid working to find trading strategies. The more people that join, the better the trading strategy. It's currently providing trading signals profiting about 60% per year, but it's expected to go higher over time. Forget about searching for aliens, let your computer turn a profit for you! www.stratagenie.com

While nobody can predict the future, you can actually predict the statistical likelihood of the stock prices over a short term. Using that knowledge, you can trade in and out while the market is behaving "normally". In the event of a rapid departure from normal, the unusual activity is seen as "volatility" and algorithms can react accordingly.
[ link to this | view in thread ]