Long story short I was reading about the ever=fascinating and slightly irritating Ray Kurzweil earlier, and per Wikipedia, apparently he began a venture in 2006 to use Artificial Intelligence to drive quantitative investment programmes, dubbed FatKat (Financial Accelerating Transactions from Kurzweil Adaptive Technologies).
What got to me though, was this:
"...he predicted in his 1999 book,
The Age of Spiritual Machines, that computers will one day prove superior to the best human financial minds at making profitable investment decisions."
Ray has a history (and a slight ego to boot) of making predictions for technological change, many of which have turned out to be both prescient and accurate. One cannot help but wonder how colored his judgment may be from his successful past experiences applying technical and scientific thought advances to other fields such as education and healthcare though.
Finance, and more specifically profitable investing, presents a variety of challenges which have brought down virtually every systematic strategy over any significant period of time, going back as far as such things have been recorded. Until the past year or two one had to go back an entire decade (whoa!) to the spectacular chronology of When Genius Failed , but over the past 24, or really 12 months we've seen the 'new breed' of quant geniuses getting crushed as badly, if not worse, than their non-quant brethren.
I think if Kurzweil, himself, cared to contribute to this rambling, his argument would hinge on the fact that the failure of quantitative investment schemes (or lack of relative success) of late is due to the human element, not necessarily the technological, although the two are inevitably and inextricably linked whether they like it or not.
The most historically spectacular investing failures, regardless of the system (or lack thereof) employed by their practitioners, have shared many a characteristic, among them ego, hubris, pride, arrogance, and intractability rank at or near the most frequent and serious offenders. My inclination, one I believe is supported repeatedly by history, is that there is not one single investing strategy which (out)performs in any and all circumstances. I realize this is not necessarily the claim of many quant funds, however, in the grand scheme of things, it seems to be the ultimate aspiration, or at least nerdy wet dream of many.
The quants' search for the ultimate algorithm, for the AI that's always one step ahead of mere mortals may very well be an impossible task, or even a fool's errand. That there is a very public, multi-party game being conducted, where each party believes that, utilizing essentially the same set of tools as the other parties, they will outsmart them, seems inherently ridiculous to me. Of course, this can also be said for the larger investing game as well, although the number and variety of approaches (not to mention the "quality" of many participants) opens up the odds some, although the point is hardly lost on me. Much in that same vein, what we've seen is that rather unsurprisingly, the proliferation of quantitative investing has created an even more profound feedback loop in various markets than already existed just with the more 'traditional' methods and participants.
The logical action for the prudent quant is then to adjust the models to better account for the actions of others with similar strategies. Throw in another factor here, tweak a little stochastic calculus there, no big deal. Of course, if everyone (or a significant # of participants controlling a significant amount of money) makes these changes, then you run into the unsurprising eventuality where everyone is essentially second-guessing each other, and thereby second-guessing themselves, much in the same basic fashion even the most novice retail brokerage account holder approaches the investing game.
What then, is the purpose, besides to tout one's technical or mathematical acumen, if the results do not stand up to scrutiny, if it can't weather the storm better than say, going to cash. Contrary to many a financial professional, being down less than your (bullshit) "benchmark" in a severely down market as we're currently in does not strike me as worthy of celebration. When, in a severely down market, you've outperformed cash, well Sir, now we're talking! Crack out the good shit, we're having ourselves a party tonight!
Now, back on track (one resembling the Nurburgring, perhaps).
As many an investor has acknowledged, in this game, being right, and being profitable aren't necessarily one-in-the-same. While admittedly comparing apples to an orange grove, over the past year or two many technically advanced quant funds, with access to a virtually unlimited supply of data and analytical power, suffered serious losses. During this same period, myself and several acquaintances, with significantly less resources (to say the least!), have achieved positive returns, all-the-time wondering how it was possible that everyone else could get it so wrong. By "it", of course, I mean the basic conceptual bets. For example, as I'd mentioned over the course of the past 9-12 months, the short Retail trade seemed completely obvious if one was willing to look at the body of evidence available. When I shorted Lululemon (LULU) at around $36 (after recommending it months earlier as a short before I started @ 1-2 Knockout when it was trading over $60), the analyst estimates were based on entirely unreasonable growth rates, given what I saw was the inevitable shitstorm we were heading into, the same shitstorm in which we currently find ourselves. I can't speak for every quant model of course, but of late it seems to me that many such funds failed to see the forest for the trees. That is, the data was there, but when one is looking for very complicated relationships, perhaps its not too surprising that some very simple connections went relatively unnoticed.
My intent is not to highlight my genius (nor to even claim such a thing exists; hint - it doesn't), but to simply point out the fact that perhaps the most highly-correlated factor or driver of investment success is not the sophistication of the system, but the ability of the manager (whether man or machine) to stay curious, and informed, and to keep an open, and humble mind.
Are quantitative strategies useless? Of course not, don't be ridiculous! They are yet another tool in the prudent investor's toolbox, which, when used "wisely" can yield spectacular results. However, just like with any other powerful tool, an over-dependence upon, or misuse of quantitative analysis can be, as we've seen, catastrophic.
Excuse the rambling, its 2am 3am for Christ's sake!
Thoughts? Discuss amongst yourselves in the comments.
