I was recently watching a report on Canada's Business News Network (BNN), which was attempting to show how “flash-trading” (i.e. giving insiders advance information on orders and prices) was “bad,” but “high-frequency trading” (i.e. allowing a computer algorithm to trade for you) was “good.”
To discuss this issue, BNN interviewed a woman named Irene Aldridge, who has written a book on this subject called High-Frequency Trading: A Practical Guide to Algorithmic Strategies and Trading Systems. Sadly, the talking heads in the North American media tend to assume that when someone has written a book on a subject that this makes them an unquestioned “expert” - with the talking head generally allowing such people to make any assertions they want unchallenged.
In this case, the author interviewed by BNN made the sweeping statement that the only difference between a trading algorithm and human trader was that the trading algorithm was faster – meaning we should all embrace the “efficiencies” these programs bring to trading. The author assured BNN that these algorithms “knew everything” that those doing the programming knew.
Ironically, one of the limitations of new technology is a lack of understanding of the limitations of new technology. We all know that a computer can be programmed to play chess as well as any human, so why can't computers handle most/all of our trading for us?
While I'm no computer programmer, my background in mathematics means I have a much better understanding of algorithms than your average person. I can state unequivocally that there are enormous differences between creating software to play chess, and creating software to trade markets.
The most important difference is that a chess game is a closed system! In a chess game, there is no possibility of any exogenous events upsetting the program through introducing unexpected variables. Conversely, in trading markets there are literally an infinite number of exogenous events which can impact markets and introduce totally unique parameters at any given moment.
From political and economic events, to extreme weather and natural disasters (and any combination of those parameters) there is an endless list of events which can occur on any given day. To make our world more complicated still, there is no way to predict how such events will be reported. If a hurricane and an important political event occur at the same time, there is no way to know in advance which news item will be stressed more in news reporting, meaning there is no way to know which of these events will have a relatively larger impact on markets.
In addition, the number of possible “moves” in a chess game are not only finite, but in terms of computers, represent a very limited number of options. Conversely, markets literally present infinite possible “moves,” and infinite possible outcomes. It is impossible (even for a computer) to make optimal decisions in a system with infinite possibilities. At best, it could make near-optimal decisions, most of the time. However, during times of extreme events (such as we experienced last year), inevitably these programs make worse decisions than a real person, most likely aggravating periods of crisis like we just experienced.
This alone is more than enough to invalidate all trading algorithms, with respect to their ability to handle atypical scenarios. The moment an algorithm is forced to perform in circumstances which are outside of its programming, these algorithms descend into chaos – where suddenly there is no way to predict how the algorithm will behave. Allowing dozens of these algorithms to function, with thousands (or tens of thousands, or hundreds of thousands) of individual traders operating them doesn't make markets more “efficient.” Instead, the algorithms themselves become unpredictable variables which do nothing but amplify volatility.
However, this is only the beginning of my objections to a system which has done nothing but automate “stupid money.” After all the market “experts” (the same “experts” who design trading algorithms) finished exclaiming how “surprised” they were when the largest asset-bubble in human history burst, they have spent an equal amount of time pointing out how “unprecedented” recent events have been. How exactly can a programmer program a mindless computer to handle situations which the programmer has never experienced in his own life?
The answer, of course, is that he can't. Thus, the trading algorithms have been woefully inadequate to handle recent conditions in markets. Rather than making markets more “efficient,” they have simply made markets more unpredictable as these algorithms blindly stumble through each trading day, making buying and selling decisions which they lack the experience and understanding to make in a rational, optimal manner.
Worse yet, these algorithms all incorporate the implicit assumption that we have “free and open markets”. As I have pointed out several times, most recently in “Insider-selling STILL accelerating,” it is ludicrous to suggest that the United States has “free markets,” when it’s common knowledge they have their own “Team” of manipulators interfering in markets every day, while a mere handful of banker/broker oligarchs literally control most of the trading. I read one estimate that 75% of all U.S. equities trading flows through Wall Street.
How do trading algorithms react to manipulated markets? I have no idea, since none of these algorithms has any programming to guide them in how it should behave. However, with both the Plunge Protection Team and the Wall Street crime syndicate being completely familiar with all these algorithms, and (as the manipulators) also having full knowledge of exactly how trading has been distorted (by themselves), it is even easier to lead these trading algorithms around 'by the nose' than the reckless lemmings known as “retail investors.”
Thus, the reason why the Wall Street crime syndicate is such an enthusiastic supporter of high-frequency trading is because it makes it even easier to “game” the markets. Additional support comes from a long list of other market participants, who blindly endorse these programs (despite having no understanding of their limitations) – simply because it is easier.
The argument in favor of trading algorithms cannot be based on “efficiency” or “practicality,” since these programs are neither efficient nor practical. Indeed, ultimately the only real arguments in favor of these stupid, computer programs are based upon greed and laziness. Quite obviously, these are considerations which do not lead to favorable outcomes for most market participants.
