'Dark Pools' is a fun, intelligent, beach weekend read.
Apart from the obvious Wall Street / HFT focus, the book struck me as a cross between War Games, The Matrix, and Terminator: Rise of the Machines.
You immediately enter a world of high tech mayhem, with super-algos, blaster bots, and hunter-sniper cloaking devices duking it out at the speed of light.
Re: War Games parallels, you get an iconoclastic, vaguely teenage anti-hero - Josh Levine, the misfit-hacker-genius creator of Island - who pursues an idealistic vision of "making markets free," with no impure thoughts of capitalistic gain, until one day his mutated ECN creation all but comes alive and says: "Good morning, Professor Falken. Do you want to play a game?"
Re: Matrix parallels, by the end of the book we have fast-forwarded from the humble beginnings of electronic trading to the near birth of AI (artificial intelligence)… the "desert of the real" (in this case the real being markets)… and the grand vision of supervillainesque networked undersea substructures, monitoring global data flows from strategic ocean points all around the globe.
Re: Terminator, in the final stretch I kept waiting for Patterson to write: "As future tech historians will note, SkyNet showed signs of self-awareness on X-X date, 2012…"
It was all a bit much - but in a good way. As Keynes once said, "Words ought to be a little wild, for they are the assault of thoughts on the unthinking." This book will definitely get you thinking about the impact of high frequency trading on markets.
My two cents: At the end of the day, high frequency traders are the new market makers… the superfast replacements for the hand-signaling floor traders and post-sitting NYSE specialists of old.bYours truly predicted as much would happen in a review of "The Predictors" by Thomas Bass, titled "The New Market Makers?" circa 2005.
As I wrote nearly seven years ago:
These guys occupy a very specific niche in the market ecosystem. Before the onslaught of computers, human floor traders provided vital liquidity to the markets (and got paid plenty well to do so). As physical exchanges lumber towards extinction, `smart' algorithms are filling the shoes of floor traders, extracting profits tick by tick with high volume, high frequency strategies. These automated players are thus becoming the new liquidity providers and market makers of the 21st century. Daytraders and scalpers may find themselves swept up in a technological arms race, but longer term traders and investors have little to fear… it's a different game.
Such is why this review is titled, "Meet the New Market Makers (Same as the Old Market Makers)". In a lot of ways, despite all the technological advancement, the biggest things haven't changed.
Take the infamous "Flash Crash" of May 2010, for example. When you understand the role that high frequency traders play these days - in terms of facilitating the majority of volume and liquidity in markets - it makes sense to expect chaos when they all "pull their bids" at once.
From the perspective of a freak occurrence, where a large portion of the HFT community "backed away," exposing ridiculously far-off placeholder bids that were never meant to be hit, the market makers of the 21st century acted just like the market makers of the last century amidst the crash of 1987. They left a void at a point of severe dislocation, just as the old school guys did so long ago.
Perhaps now that GETCO - which stands for Global Electronic Trading Co, the most supervillain-like shop of them all - has assumed official market maker duties in many household names, such bid pulling will not be a future problem.
The question remains: Are the new guys worse than the old guys? I'm skeptical. Anyone nostalgic for the old days of physical pits and human specialists may not remember the day-to-day reality of such a system. As an international commodity broker in the late 1990s, I had the privilege of phoning into the pits on behalf of hedge and commercial clients, to yell at some guy named Vinny or Frankie or Sol - inevitably the brother-in-law or cousin of the Refco floor trader who executed our order - to try and get restitution on a criminally bad fill. This kind of thing happened far too often.
And as for being an NYSE specialist? Talk about a license to print money. There is a reason such jobs were handed down from one generation to the next. In many cases, the opportunities provided were the legal equivalent of stealing.
Not to mention the commissions - good lord, the commissions (!) that retail and institutional clients alike were forced to pay in the old days. Add it all up and I don't think the pennies and nickels hoovered up by the HFT shops, mitigated by the incredibly low-cost commissions available via technology today, amount to such a bad deal.
A bit of wildness that made me laugh out loud was the notion that computers are going to take over the markets one day, as in, putting directional investors and traders out of a job. Seriously? Puhleeze. These bots may be great at nano-scalping, playing for blips on a mass scale, but true directional market involvement is another matter entirely.
If you're truly worried about thinking machines eating your lunch in a multi-day or multi-week time frame, don't be. Marvin Minsky, a noted forefather in the Artificial Intelligence field - a very confident AI optimist 15 to 20 years ago - recently admitted the following:
The bottom line is that we really haven't progressed too far toward a truly intelligent machine. We have collections of dumb specialists in small domains; the true majesty of general intelligence still awaits our attack.
"Dumb specialists in small domains" well describes the proliferation of tick-hungry algos. They are good at what they do in a very tight timeframe, but the inputs required to parse incalculable variables across extended time horizons are another matter entirely. As I wrote some time ago, the most powerful supercomputer on the planet is not smart enough to figure out the turbulence in a glass of water - and yet we expect it to crack the self-referential human feedback loop that is markets?
The book closes with a glimpse of the supposed future in "Star," the self-learning, self-teaching virtual machine assigned to make investment decisions for a tiny hedge fund, Rebellion Research. To the extent that Star is supposed to be a threat to humans, color me skeptical. (Those who are just mediocre at their jobs - rather than very good - have much to feel threatened by, of course; but such has it always been.)
Patterson makes brief allusion to computer-assisted chess, the powerful combination of hardware and wetware (software programs plus human guidance). For cream of the crop money managers, I think this is closer to the true way forward - using technology to enhance human capability, not replace it.
In terms of thinking games, the best metaphor / AI-intelligence test is perhaps not the fixed Western game, Chess, but the fluid Chinese game, Go, which contains far too many variables within the scope of possible movements for any computer, even Deep Blue, to brute-force calculate the optimal strategy path. (For this reason, no Go champion has ever fallen to a machine.)
At the end of the day, trading and investing strategies will continue to evolve. High frequency traders, and various forms of new technology, will continue to influence markets in unexpected and interesting ways. But I believe the following words from "Reminiscences of a Stock Operator," which were true in 1923, will remain just as true a century hence:
There are men whose gait is far quicker than the mob's. They are bound to lead - no matter how much the mob changes.
Smiling at the bots...
Disclosure: As active traders, authors may have positions long or short in any securities mentioned. Full disclaimer can be found here.