Given everything going on in the US today (and the fact that I can’t tear myself away from CNN…damn you Wolf Blitzer), today doesn’t feel right for a data-centric post.
So I want to add a little bit to the blogosphere’s “WTF happened in October” file with my take on why so many active stock market trading strategies failed last month.
From the perspective of a guy on the “inside” of strategy development (i.e someone who is a developer, active in the developer community, and works with a lot of new developers), I would guesstimate that over 90% of active trading strategies either took it on the chin or were completely knocked out in October.
“Of course”, you might think, the market was just that ugly. But think about that for a moment. If active traders were really employing a broad range of strategies (contrarian, momentum, trend-following, etc.) we would expect to see a far wider range of winners and losers.
But the fact that most folks don’t like to admit is that the overwhelming majority of active equity trading strategies employed are contrarian. By “contrarian” I mean they are buying things that are weak and selling things that are strong. There are a lot of flavors of this candy: different time frames, different assets, different indicators, etc. but at their core, most share this common element.
Why? Because historically, contrarian approaches have (and I believe will continue to) work. I owe my own personal investment successes to my own flavor of this candy.
But therein laid a problem. October was every contrarian strategy’s worst-case scenario: a market that moved in a single direction in a way that busted all historical norms (in the same vein though still very different than October 1987).
That last bit about busting historical norms is key. A good strategy, regardless of its approach, should be built to handle what we’ve seen before. But when the market acts in a way we haven’t seen before – an outlier, a black swan, a fat-tail - all bets are off.
A side note: does anyone else think it would funny to apply yo mama jokes to fat-tails as in “yo tail is so fat it lost a game of hide and seek cause I spotted it…behind Mt. Everest”. Anyone? Just me? Ok…I digress.
So there you have it…a recipe for a very bad month for active strategies. The vast majority of active strategies are contrarian because historically that has worked. And in October, the market maximized contrarians’ pain.
A TALE OF POSSIBLY-ILLEGAL ACTIVITY:
A few years back, a friend of mine (who shall remain nameless) discovered a flaw in the website of an extremely respected company on the Net (who also shall remain nameless) that offered strategy development software. This flaw (which has since been fixed) allowed said friend access to over ten thousand very private strategies developed by thousands of other developers – from the amateur to the professional.
Said friend, being a good bit sneakier than you or I, downloaded these thousands of strategies and then built an engine to test all of them over all sorts of assets. What did he find? Over 99% of the strategies that showed an ability to consistently outperform the market were some flavor of contrarian candy.
If that’s not evidence of the fact that active traders are dancing variations of the same jig, I don’t know what is. There are still good and bad strategies, but the majority are still vulnerable to the same inherent threats.
Happy Trading,
ms
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Random Stuff | 10 Comments
Interesting perspective. I think might be right. Maybe buying cheap stocks and tossing away strong ones is more psychologically appealing to us. Everyone loves the puppy, but grows tired of the dog.
ms,
Interesting.
I actually have a mate in Oz who went the other way with a mechanical system, I’ll post the code for you later if you are interested.
But in keeping with your theme, a number of Oz developers all went the same way, trend following, all on the ASX.
This might be because the ASX was a bull trend for a number of years. None of their systems worked on the US markets however.
jog on
duc
Mike,
You’re absolutely correct, and I’d bet anyone who’s been actively developing systems for any amount of time has arrived at the same conclusion (contrarian trading outperforms trying to catch trends). I call it “Swing Trading as the Market breathes In and Out”. And (those of us who’ve arrived at this conclusion) have seen in our records how the decent-to-good gains these happily-swinging pendulums produce would have been spectacular (in some years, unbelievable) except for a few large drawdowns…. those times when we get an inhale, take a position, and then the market holds its breath, going the same direction farther and farther until finally it settles and starts breathing again. It’s those 1-4 times a year which account for only a small minority of our losing trades, but for the vast majority of our losses.
My last couple of years have been spent about 25% on system development and about 75% on searching for the Secret Ingredient, the one which, if nothing else, would simply cash out and stop the system from trading while the market is holding its breath, thereby avoiding a decent portion of each of those large losses. (I think we’ve all tried and proved that virtually no simple Stop or Trailing-Stop strategy seems to achieve this goal, they usually make things worse). This Secret Ingredient would only kick in maybe once every few months, but it would vastly improve the profitability of the system (virtually ANY of the contrarian systems).
What caught my eye was your exploration of “A New Approach for Coping with Abnormal Markets”. It occurred to me that you too are on the trail of the Secret Ingredient. I honestly think it’s a search than can potentially make a difference of multiple decimal places in future returns.
This past July, I found my own Secret Ingredient (first edition, at least), and while it still has plenty of room for refinement, it pulled the plug on trading and went to cash the second week of September, and has now (early Nov) given the signal to “resume trading”. It ain’t perfect, as they say, but I’m quite pleased.
Wait, that’s not really why I started this comment. What I meant to say is this: Notice how all of the backtesting strategy stuff has a significant online component. Y’know how the big brokers actually have systems to trade against their crowd of inexperienced traders’ orders (one of the most profitable strategies ever, no doubt). I have no way of proving it, but I’d bet you my big toe that, if their organizations have any brains at all, they have systems set up to skim the cream off the top of all the millions of backtesting strategies run through their servers daily. Their own version of Amazon’s Mechanical Turk, with thousands of (us) individuals toiling away each night, providing the free labor. Not that there’s anything we could or should do about it, but just a thought.
Anyway, cheers and best of luck, love your work and have learned much from it.
Things come and go and come back. Contrarian has ( had? ) its day, so did momemtum. They stop working, they will come back.
Remember the Turtles ? Those were the days of momemtum (=non contrarian) trading. Are they gone forever ? I doubt it.
To everything there is a season . . .
Perhaps it’s a time frame issue. This study concludes that strong stocks are persistent.
http://www.blackstarfunds.com/files/Does_trendfollowing_work_on_stocks.pdf
RE to duc: good point…I should qualify my post by saying “in the US”. I work almost exclusively with timers in the US market, but the little bit I’ve done on other markets does say that some markets move to a different drum (some respond to momentum indicators, not contrarian indicators). I should also have stressed that I’m talking about US equity markets. Other assets (currencies and commodities come to mind) are much more trend following than equities.
Thanks for the comments duc.
michael
RE to Will: thanks for the thoughts – good to hear from someone who obviously has put a lot of sweat equity into mechanical systems. I only have one thing to add re: what you call the “secret ingredient” that moves to cash when the market isn’t going to respond to contrarian indicators. You’re right, my approach (detailed here: http://marketsci.wordpress.com/2008/10/15/a-new-approach-for-coping-with-abnormal-markets-shades-of-grey/) attempts to do just that. I think the biggest difficulty in building something like this is that these events are, by their nature, few and far between (and difficult to predict). I think it’s very easy to get fooled by curve-fitting. That’s why I took a scaled approach rather than a binary (on or off) approach. Certainly not foolproof and not curve-fit-free, but a bit more robust.
Thanks again for all the thoughts.
michael
RE to eber: agreed. I don’t think it points to a “fad” – I think it points to the actual mechanics of the market evolving (ex. http://marketsci.wordpress.com/2008/07/17/evolving-markets-dynamic-systems-daily-follow-through/). The problem is that I think most people are very static in their thinking and won’t be prepared when the next shift comes (and I agree, it will come).
michael
Michael,
I find your blog very stimulating, I’m glad you share your ideas with us.
Lately I have also been using almost exclusively contrarian strategies, trading individual stocks in the US. The experience from these past two months also triggered me to want to refine/supplement some of these strategies. Compared to August 31 my account is actually up, but in between I experienced some pretty steep drawdowns, which I found to be too stressful. So I started investigation in two directions:
1. A combination of medium term trend following and short term swing trading strategies. This will likely reduce drawdowns, while possibly also lowering returns, since part of the trading capital would not any more be available for contrarian trading. But I consider this as some kind of an insurance.
2. My most profitable swing trading strategies employ a time frame of a few days. However, I have some strategies with a time frame of 3-6 weeks, which can still generate attractive returns, but they are not nearly as powerful as short term trading. So I have not been putting much money into these strategies. But now I realize that combining the strategies of the two time frames actually would have smoothed performance in the past, and especially this year, while returns would not have been hurt in any substantial degree.
Kind regards from Hungary
RE to Andras: good comment and you and I are thinking some of the same things right. I have been considering working towards building a more long-term strategy employing more of a momentum/trend-following approach. I think you’re right that in terms of pure return a good short-term strategy wins the day, but I think coupling that with a long-term strategy would have two benefits: (1) potentially reduce portfolio vol (because it’s a different radically approach to trading the markets), and (2) in my case, a number of my investors/subscribers are looking for something a bit less short-term as what we’re doing now so that would allow us to fill that gap.
This would never be a huge component of what I do because I think over the long term it would reduce returns, but throwing say 25% of my portfolio towards it I think would be reasonable.
Thanks for the thoughts,
michael