Thursday, December 3, 2009

Geeks Trump Alpha Males as High Frequency Trading Takes Over

For those newer to the website, we have been early pointing out the changing character of the market, due to the increasing dominance of computer based trading.  I first started noticing it in 2006, when the market was wickedly "calm" if you will... this was pre blog and I had been noting on some message boards how the S&P 500 had not dropped more than 1% in any 1 session for more than 10 months (I don't remember the exact figures but it was some absurd statistic).  I began wondering about computer based trading and the increasing dominance of ETFs as "monolithic trading" began to really enter my conscience.  The incredible performance of James Simons of Renassiance Technologies had also stirred by imagination... what the heck was this guy and his team doing that was minting money year after year? [Apr 8: Hedge Fund Manager - Good Work if you can Get It]  Effectively, I believe much of what has happened the past 3-4 years is a massive "copy cat" of the Simons model.

Upon starting the blog a year later, the events in August 2007 really caught my eye as very strange behavior - later identified in a New York Times piece (somewhere in our 5000+ posts on the blog) to be largely due countless algorithms all plying the same trades but being shaken by dislocations in the market - caused incredibly strange behavior (massive volatility) in some select group of stocks, some of which I owned.  We began calling this HAL9000 on the website... when my audience was a whopping 100 page views a day... and now I see "HAL" references on countless blogs.

If you are unfamiliar with what has happened, essentially (based on whom you ask) anywhere from 40-60%+ of all trading now is automated, and a good portion of that is now done in millisecond increments.  Overlaid with the dominance of ETFs, both for indexes and specific sectors, this has changed the character of a market in a very tangible manner.  It is also why I believe technical analysis has come to dominate fundamentals increasingly over the past half decade.  Over the past 18 months I am seeing a lot of people acknowledge this...[Jul 18, 2009: Joe Saluzzi Comments on HAL9000] [Aug 6, 2008: Cramer - Quants and their Machines]

What is truly amazing is the human leverage... a firm like Getco with a mere 250 people controls 10 to 20% in all trading in many US stocks.  Think about that for a moment.... it is stunning.  [Aug 28, 2009: Meet Getco, High Frequency Trading King]  Getco more powerful than Vanguard or Fidelity... combined.  250 humans.  Can you imagine what Goldman can do with 500 people? 750?  We'll never know... that I am sure of.  All I know is Bernie Madoff was on the loose, despite well over a decade of warnings ... but the minute some Goldman Sachs algo code was (allegedly) stolen the FBI was on the case within hours.  If it's that important to markets you have to wonder, how powerful is this stuff?

Eventually, like all "great ideas", this automation will first (a) get over crowded and then (b) cause some sort of 'Black Swan' event... the fearful thing is with so many interconnections & the speed of these computer actions, very foreboding things could happen literally in minutes.  I am assured this could never happen, just as I was assured by the experts in 2007 that housing prices could never fall on a nationwide scale.  And that Fannie and Freddie are just fine.  And... well, you get the idea.  Actually, it is sort of funny - now that the market is going in the "correct" direction, all the concerns about HFT seemingly have disappeared.   Funny that.

We won't even get into the talking point we've brought up in the past about long term impact on a society when many of its greatest minds are being use to game theory millisecond stock movements, (paper shuffling) rather than in research & development, science, engineering, or new business creation.  But that is the way the compensation system is set up, and talent will go where the money is - can't blame anyone for this.

One last point before we get to the stories - something I have been thinking about the past 6 months as I watch every trader I know chasing after the exact same technical breakouts, and knowing every computer is also set to these same exact conditions.  (such as yesterday and today, buying a "breakout" over S&P 1112)   Say you had a power player in your administration who really wanted to get the US markets up to increase confidance and "good feelings".  And he had experience working in a computer dominated hedge fund shop to understand how the electronic markets work behind the scenes.  And he had access to virtually unlimited amounts of US treasure to buy S&P futures at appropriate times to goose the market each time it looked technically weak... or when it needs a technical breakout.  It would almost be like the Pied Piper leading markets around by his magical flute.  Well, that would be one powerful man... especially when he is considered one of the brightest people in Washington D.C. [Apr 6, 2009: Larry Summers - No Conflict of Interest; He Pinkie Swears]  Hold on... someone is yelling at me to get off the grassy knoll.


As I was surfing Reuters last night, I was struck how 2 of the top 10 stories were both about high frequency / algorithm based trading - thought they were worth sharing.  It seems not only is the nature of "investing" changing, but the type of person behind the scenes is slowly morphing as well.   The "geek" shall eventually inherit the Earth stock market? 

1) Reuters: Geeks Trump Alpha Males as Algos Dominate Wall Street
  • Wall Street traders aren't what they used to be -- they're not even on Wall Street anymore. The days of swashbuckling backslappers on the floor of the New York Stock Exchange have given way to an era of trading dominated by analytical technical whizzes whose computers may be running from a town in deepest New Jersey or Texas.
  • While street smarts and an ability to socialize were crucial to successful floor traders, today's trader needs math and computer science, and quite possibly a PhD.  And that has to be coupled with coolness, organization and logic to sift through masses of trading data each day and think about how to shave microseconds off trades.
  • The outsized growth of high-frequency trading, dark pools of liquidity and high-tech computer algorithms has fundamentally changed the game on Wall Street -- and the psychology of those who work there.
  • Traditional floor trading "really is an alpha-male activity," ..... "You get these highly competitive people taking a good amount of risk ... It's like being in a locker room. In contrast, computer programmers are almost like a think tank."High-frequency traders are practical, problem-solving people with an engineering background.
  • Now, with high-frequency trading representing some 60 percent of U.S. stock trades, the atmosphere appears to owe as much to Arthur C. Clarke and artificial intelligence as to Gordon Gekko and the 1987 movie "Wall Street.  "They are introverts, some are socially awkward, and they don't seek publicity. They are the type of guys you would see at a Star Wars convention," said Sang Lee of Aite Group.
  • "These are not your Harvard B-school grads, per se," said Robert Olman, president of Alpha Search Advisory Partners, an executive search firm for hedge funds and proprietary trading shops. "They often have dual degrees, bachelors and masters. One degree is going to be computer science, and the other degree might be financial engineering, math, physics," he said.
  • And it can be very lucrative, with a programer typically making 10 percent commission on the money his model generates, said Irene Aldridge of Able Alpha Trading, a one-time quantitative specialist at CIBC in Toronto.  The best programmers can make tens of millions of dollars a year. That was even the case during last year's financial crisis, as great volatility offered both risks and opportunities for high-frequency traders.
  • "It's a highly technical, mathematical game," Berkeley said. "They are playing a very precise game of statistically estimating and predicting over the next three to five seconds whether there is going to be any liquidity in that stock and where it is. And how they can take it without being seen and without leaving any tracks."  (uhh, sort of like Benjamin Graham - yeh)
  • Low key and somewhat awkward, these introverted but brilliant traders look up to James Simons, the Renaissance Technologies fund manager known as the "King of the Quants."   A media-shy mathematics professor, the 71-year-old Simons has made billions of dollars by making the right bets on technical trading strategies. In January he will retire from his firm, which manages $17 billion in assets, but will leave an indelible mark on the industry.
Location, location, location
  • An obsession with milliseconds has led high-speed traders to focus on "co-location," where trading shops try to place themselves as close as possible to an exchange's data centers.  High-frequency trading shops that focus on options have sprung up in Chicago, near the options exchanges, and in New Jersey, where trading venues like BATS and Direct Edge are located and where the Nasdaq and New York Stock Exchange house their data centers.
  • "It's just like having a high-speed fighter airplane versus a slow-speed one. If I can turn inside your turning radius, and if I'm faster than you are, I can win every time."

And as I noted above... of course nothing will go wrong, with trigger happy computers all dueling each other in time periods shorter than 100 blinks of an eye.  The "waterfall" effect never causes issues, does it?
  • However, high-frequency traders' computer programs are not foolproof.  John Malitzis, vice president of market surveillance at the New York Stock Exchange's oversight body, told a conference in September that he has seen examples, both in U.S. markets and elsewhere, of computer algorithms malfunctioning. He calls them "algos gone wild."
  • Proprietary trading shops and banks spend millions developing complex algorithm strategies and want to keep them in a black box well away from their competitors.


2) If you are interested in the topic, Reuters has an even more in depth article titled "Who's Afraid of High Frequency Trading?" here.  Again, I cannot stress if you are trying to figure out how the market "works" nowadays, reading old value investing text (I believe) is becoming increasingly useless. 

The market "is not trading on fundamentals anymore. It makes no sense, it's very frustrating for traders," said Alan Valdes, director of floor trading at NYSE member Kabrik Trading. "It's all programs."

However, understanding HAL9000 and his friends dominant stature is so very important.  

Institutional investors mostly complain about alleged unfair advantages and that their trades are being "gamed."  A high cancellation rate for orders has sparked suggestions that the algorithms are deployed to glean information from pending order flows, and then based on that knowledge, race ahead to scoop up trades.

More than 90 percent of orders submitted to the New York Stock Exchange by high-frequency firms are canceled.

The head trader of a European money manager with more than $100 billion in assets said high-frequency traders profit through pattern recognition software to anticipate a trade.  "It's a certain knowledge of what's coming. It's not like they're guessing what's going to happen, they're not speculating," said the trader, who spoke on the condition of anonymity. And he added emphatically: "They know."

And you wonder why Goldman Sachs (or RenTech) is just so much "smarter" than everyone else in the room?

[Oct 1, 2009: High Frequency Trading on 'The Daily Show']
[Sep 4, 2009: Meet Optiver, High Frequency Trading Friend in Oil]

Disclaimer: The opinions listed on this blog are for educational purpose only. You should do your own research before making any decisions.
This blog, its affiliates, partners or authors are not responsible or liable for any misstatements and/or losses you might sustain from the content provided.

Copyright @2012