Slashdot Mirror
Quant AI Picks Stocks Better Than Humans
Mr_Blank writes with this excerpt from an article at MIT's Technology Review:
"The ability to predict the stock market is, as any Wall Street quantitative trader (or quant) will tell you, a license to print money. So it should be of no small interest to anyone who likes money that a new system that works in a radically different way than previous automated trading schemes appears to be able to beat Wall Street's best quantitative mutual funds at their own game. It's called the Arizona Financial Text system, or AZFinText, and it works by ingesting large quantities of financial news stories (in initial tests, from Yahoo Finance) along with minute-by-minute stock price data, and then using the former to figure out how to predict the latter. Then it buys, or shorts, every stock it believes will move more than 1% of its current price in the next 20 minutes — and it never holds a stock for longer."
It's been said before, and I'll say it again. You should ban anyone buying a stock and then selling it within timeframe x (where is 1 week/6 months/1 year). Anything to cut down on the insane bullshit.
Nothing could possibly go wrong!
Conservation of angular momentum makes the world go round.
and it never holds a stock for longer
So this is really an automated gambling system rather than a tool for investment.
I am becoming gerund, destroyer of verbs.
This type of day-trading provides absolutely no value to the economy and should be regulated to death. Also, I have a hard time believing it would work in the long term, as the news is a lagging indicator 90+% of the time. It might work for shorting in that respect... but even that would be too late and high risk.
All these quant systems seem to do is increase volatility at the expense of the market establishing a general direction.
I'm all for an IRS withholding of 1% on sales of assets held less than a week, and I am a fairly active trader.
Ars Technica wrote an interesting article about this almost a year ago. What is happening now isn't anything all that new. As several people have already mentioned, yes this is dangerous because these tools trade in extremely large sums. Slashdot even covered United Airlines stock dropping from $12 to $3 when the news crawler for one of these tools thought an old story was new and the tool proceeded to dump its entire United holdings causing a massive sell off by other investors. http://arstechnica.com/tech-policy/news/2009/07/-it-sounds-like-something.ars http://tech.slashdot.org/tech/08/09/10/203233.shtml
It doesn't. I don't see anybody claiming it does. It makes money for those who created the system. Might not be the most noble goal, but it's sure as hell a sensible one to go for.
We do not need to remove them. Regulations keep the playing field more even. Without them we'd see most of the wealth of this country flow up to a select few.
Blar.
You know what a long term investment is ? a short term investment that failed...
It takes 40+ muscles to frown, but only four to extend your arm and bitchslap the motherfucker
A trade every 20 minutes would generate an obscene amount of transaction cost. These are the costs associated with stock trading and any successful Trader/Manager team would seek to minimize these things and create a strategy that maximizes gain for the least amount of trades. Given this system is trading so frequently it would seem to assume zero transaction cost. Quite
I would also think that any more than a few of these trading bots would create a market impact that would nullify the advantage eventually.
Its also kind of weird to invest based on information that is not reliable, or not stemming from the fundamentals of a stocks earning potential. Creating portfolios based on glamour, popularity or essentially inefficiency is a disaster waiting to happen. These gimmiks will work until a catastrophe happens, but by then so much more is committed than is was ever prudent and the damage is done. Investing based on artifacts or involving anything you don't understand with your own 2 hemispheres means that you as an individual are being manipulated and you are investing for irrational reasons.
Unless this technique can be balanced or controlled in a risk management context, or understood at a low level to behave in parallel with existing market benchmarks, then only the penultimate fool would see this as responsible investing and not patent gambling and recreation with your surplus funds.
While I see the expectation of any product to magically increase in value over time as fundamentally insane, there are securities that can offer utility in markets. If this kind of information is really that valuable to a majority then it will eventually be securitized. If this is something that only a minority can exploit and is not private information, some Tony Soprano somewhere in the world will call it illegal. And thats the other reason not to bother - because none of the markets are truly open and will never be truly efficient while goverments are waiting to intervene. Maybe there is opportunity in the many small irrationalities of the human mind since the market participants are almost all human. What else could this information harvesting be capturing besides some departure from fundamental consideration? If the fundamentals are considered and reflected in the price, then any other change is short term and technical, to be nullified.
I should watch what I say lest I bump up Apple stock!
There is an alternatively and it just might address the capital gains issue. We tax capital gains at the same rate no matter how long you keep the investment. Why not have a sliding scale that bases the capital gains tax rate on how long you hold on to the stock. Suggested tax rates. At the same time, investors are always crying that capital rates are too high. With this scheme, they would be in control of what rate their investments would be taxed at.
1 second ... 99% ... 95% ... 90% ... 75% ... 50% ... 35% ... 20% ... 10%
1 minute
1 hour
1 day
1 week
1 month
1 year
5+ years
You could even put the tax rates on a continuous scale that negates any advantage to holding on to an investment just long enough to meet a benchmark. Yes, short term investments would be taxed at a confiscatory rate, but that is the general idea. We want to slow down the rate of trading. At the same time, investors are always crying that the capital gains rate is too high. This would put them in control of what tax rate their investments is taxed at. All they need to do is to hold on to investments long enough. This scheme would also favor the little guy who probably holds on to investments for a longer time.
http://www.robschumaker.com/publications/IEEE%20Computer%20-%20A%20Discrete%20Stock%20Price%20Prediction%20Engine%20Based%20on%20Financial%20News.pdf
If this program worked at all, I'm curious why Schumaker and Chen did not keep this to themselves and just retire to the level of Sergy and Brin. With a small wad of cash to get the ball rolling, it would just snow ball into ever-increasing cash reserves.
They could start a lot of business ventures from the 1/2 acre deck of the 800' mega yacht anchored in the Med, all while their secret stock market money machine worked its wonders.
Ah, right. Web Bot makes a (more or less) triumphant return, does it?
Well, a couple of cherry-picked results (you know, such as "data from five non-consecutive weeks in 2005") could net a snakeoil salesman a couple of million, so I guess it does what it was written for.
I've seen similar things about people choosing stock at random and performing better than your average stock fonds. Some of them, anyway, pretty much what you'd expect by chance - the pros don't have a magic wand to show them what stocks will rise, they gamble just as much as ordinary folk who stumble into the stock market. The difference is that they're not gambling with their own money, and they get paid handsomely for it. After all, "professional" only means that you're getting paid for it, not that you're any good at what you do.
As such, color me unimpressed if someone finds a hand-picked dataset for a bayes filter that does better than a professional trader.
and this development helps humanity, how?
The only possible benefit I see is to help to bring the end to this fiasco that is Wall Street.
There are HFs using this strategy now using dedicated reuters feeds and trading in microseconds. This means new information is impounded into stock prices well within a second. In the article they used yahoo news and minute by minute stock data? That's laughable. I suspect the reason for their returns is that they they are indexing the time information arrived, and the price you could trade at that instant incorrectly. In other words the information arrived at t + 5 seconds, and they execute the trade at the quotes available at time t. Also I suppose they are not including margin and transaction costs, reasonable slippage, and risk-adjusting their returns?
When the trucker moves the widget from the factory to the store, he changes its value by moving it from the place of creation to the place of use. Any student of economics knows that major economic leaps have taken place when the costs of this have reduced - from carts to canals, from canals to railways. This is because there are real costs involved; you can regard the energy and investment in moving goods as being exactly as much part of their manufacture as pressing or welding. But the electronic transfer of the stock market transfers ownership at negligible cost and therefore adds no value, so any price increase is simple inflation.
This is exactly what has happened to the economy: house prices inflated, share prices inflated, but the actual value of the underlying assets barely increased. We are now trying to reduce a debt which is purely the difference between perceived value (what people will buy things for) and their inflated value.
The fact that people like you believe the nonsense you have posted is the underlying fact behind the financial crisis.
From scarped cliff or quarried stone she cries "A thousand types are gone, I care for nothing, no not one."
Originally, the stock market was not a form of gambling but a form of insurance. Investors in trade voyages in the days of sail and marine anarchy expected that some ships would not come back, therefore they wanted to be able to invest in multiple voyages. Joint stock companies formed to carry out a voyage would then sell shares, spreading the risk. (They did this at Lloyd's Coffee House in London.) The sale of shares meant that the money they had invested in the voyage came back to them before the voyage was complete, thus creating liquidity (i.e. the joint stock owners had cash again to invest in new voyages before the first ones returned).
Short term trading is purely gambling, but does not necessarily create any more liquidity than long term investment. Hence my observation that your comment is bogus.
From scarped cliff or quarried stone she cries "A thousand types are gone, I care for nothing, no not one."
Also not as an argument but just a note: the term you are looking for is Fiat money, paper money originally was just a reciept for gold and you can have fiat coinage as well.
"Fiat money"? Now that's interesting. If we had a monetary system where each unit was worth, say, 1/10000 of the cost of manufacturing an automobile, what impact might that have?
Microsoft's monopoly appears to have become limited to home and small business PC operating systems. What is Microsoft's market share in video game consoles? In MP3 players? In smartphone operating systems? In server operating systems? In web browsers, even among users of its own desktop operating system? The widely publicized antitrust lawsuit covered the use of a desktop monopoly to build a web browser monopoly. But now, Internet Explorer is right on the verge of losing its majority, and though it will still enjoy a plurality for the foreseeable future, plurality without majority implies lack of monopoly.
Close, but no cigar.
That 0.01 is NOT taken from somewhere else. It is "generated" but generated from nothing. It is the air in the bubble and then it bursts. We had this long before, the great depression was build on it. EVERYONE speculated. And LOTS of money was being generated it seemed, but where did it come from? Nowhere.
You might have heard of the phrase "your ship coming in". Where does it come from?
At least in part from the old dutch practice of funding the sailing of a cargo ship by writing out shares. Anyone could fund some money to build/outfit a ship and would in turn get a share of the profits it would generate on its voyage. This was a long term investment as a voyage to the far east could take 2 or more years. It was a also risky, you could build a bridge to the far east out of all the lost ships (oh okay, you can't but it sounds dramatic).
Now say that I took a share of 100 florins (a shitload of money but a nice round number to work with). I watch the ship sail and hope that it will come back in 2 years time with a fat cargo of spices that will trade for a fortune. My ship will have come in. Or it will sink.
BUT this ship does not exist in a void. It will encounter other vessels. Say that six months out it has crossed the horn of Africa. A seriously risky part of the journey. It comes across another ship making its way back and this ship reports what has happened to Holland. What happens to my share? Well nothing EXCEPT that SOMEONE might be willing to pay me more then 100 florins for it because the risk of it failing has now been reduced. My share has increased in POTENTIAL value. Someone with 200 florins might buy my share. I get a lower but certain profit while that person will gain less of a profit IF the ship comes in but has a higher chance of it then I did.
Other factors can add or substract from this. Say that it has been a calm year at sea and I get news that dozens of ships are making their way back. The price of spices will fall. Less risk of no return but less profit. Or say that nobody has yet reported on my ship at all. Risk has sky rocketed that it has sunk and my share is without value. Might I sell it lower?
THAT is stock market speculation. Betting on the POTENTIAL value of something. The problem is NOT with the speculation itself. The problem is when the speculation starts to be based on nothing. Those ships need to build, to be sailing, to be buying and selling cargo in order for there to be anything to speculate on. And that seems to get forgotten.
The speculation is no longer about the chances of the ship making it with a good cargo but on the speculation itself. Speculators no longer follow the shipping news but share prices themselves.
Take the recent price drop of BP shares. Why? Because of the oil spill? The company makes 60+ million profit PER DAY! The cost of the oil spill are spare change. Yes it will hurt their bottom line a bit but it is really just the cost of doing business. There should be no selling going on because the company is at no risk. Without speculators, there would be no selling going on. No long-term investor would have a reaosn to sell. Not buy perhaps but not sell since selling when a stock is going down means you are loosing money. Only the short term speculators have to sell because they can't afford to simply wait out their investment and need their money now.
We have allowed the stock market and the banks to turn themselves into "THE economy". A bank should be a service provider that real business makes use of. The same a law firm or cleaning company. Instead they have come to think of themselves as the most important part, the very engine of the economy. It is silly.
Imagine this. A justice system is part of civilization right? But when you consider the justice system to BE civilization, I think you would not like the results.
Don't confuse the means with the end. The tool with the goal.
There is nothing wrong with speculation, there is something wrong with
MMO Quests are like orgasms:
You may solo them, I prefer them in a group.
...charge a normal goods sales tax on the transaction. Why should stock sales be different from selling anything else? That will slam the kabosh on these fast computer trades.