Drawdown Blues

Strategies might be attractive on the surface but fail moving forward. Why? Here we discuss the qualities of effective trading strategies, and the many traps one should avoid.

Drawdown Blues

Postby Dacamic » Thu Apr 13, 2006 12:10 pm

My trading screen looked like someone spilled red ink a couple days ago. Every position was down, a few down a lot. The U.S. equities market struggled with high oil prices, extended valuations, rising interest rates and Iranian nuclear declarations. Frankly, Scarlet, I didn’t give a damn. This was another down day in a string of down days, and by day’s end my portfolio had staggered to a historically large drawdown. Understanding the overall market environment offered little comfort.

Wanna guess how excited I was to commit capital to new positions? Suffice it to say the seeds of doubt had found fertile soil. To confirm my pessimism, I had the “pleasure” of watching a new long position drop 4% almost immediately after my order was filled. It was kinda fascinating, really … in a car crash sorta way.

We accept losing streaks and drawdowns as statistical reality. They nonetheless levy an emotional toll. A down day can be like sitting in a seat that is being kicked from behind. An occasional tap might be annoying, but its no big deal. If jolts increase in frequency, duration and severity, happy thoughts fade … and the odds favor irrational, unfortunate responses.

Analogies, metaphors and anecdotes abound in these circumstances. Similarly, our trading colleagues may advise ideas outnumbered only by mythical cures for hiccups. My two-bit advice is not very glamorous: let your system solve its own problems. When my system’s performance statistics are consistent with back test results, I’m comfortable letting it battle its way out of a slump. Okay, maybe “comfortable” overstates the evenness of my emotional keel. Losing streaks can appear immortal, making us believe they will never end. By themselves, though, drawdowns do not justify taking a system offline.

Certainly, drawdowns are a symptom of failing systems, and red flags are expected to pop up when equity balances shrink. Warning bells should not be ignored, yet we need to consider action while we are anchored by rationale thought. A truth of our human condition is that critical thinking is often difficult when it is most important. In our trading businesses, we can better handle this paradox by monitoring our systems’ performance against whichever valid benchmarks we choose. In my case, the recent drawdown – although agonizing at times – was less than those during back tests and not significantly greater than prior actual drawdowns. With the benefit of these statistical perspectives, I was able to calm my emotions and continue entering new positions.

Even with a more stable mood, humor suffered when my next foray created the “4% car crash” debacle described above. By the close, the offending position finished up 5% for the day … ya’ gotta luv it when that happens.
Dacamic
 
Posts: 457
Joined: Wed Nov 30, 2005 12:40 pm

Postby Overload » Fri Apr 14, 2006 8:19 am

Nothing will make you question your trading system more than a drawdown phase. But your comments toward the end are exactly right -- that your trading system is operating within normal parameters based on the drawdowns it received during back testing. The markets go up and down, and trading system performance goes up and down. But as long as those ups and downs are within the range of what's been tested, you're probably safe to assume that things are operating normally and you should just wait it out.

But... I'd also add that as soon as those drawdowns expand beyond the range of your back tests, it's time to look very closely at what's going on. In fact, you may want to run some additional back tests to see just what the possibilities are in alternate time periods or using different sectors. While everything may be just fine, there's also the chance that your system has moved into uncharted territory. And we all know that means: unpredictable results.

Pete
Overload
 
Posts: 2248
Joined: Wed Nov 30, 2005 12:14 pm


Return to Curve-Fitting and Other Pitfalls

cron