Archive for July, 2009
This article first appeared in the September 2008 edition of Trader Monthly. I wrote it and it was edited by Rich Blake.
Adherence to his system enabled commodities veteran Bill Dunn to stage a monster comeback.
If you’re a trend follower, and you trade commodities, the past few years have been fun. But for Bill Dunn, a futures trader since the days of Richard Nixon’s price controls, the recent commodities explosion has been anything but a joyride.
While most trend following CTAs took elephantine profits in recent years, DUNN Capital Management, which runs $360 million, lagged sorely behind. In 2003, 2004, and 2005, Dunn’s World Monetary & Agriculture program lost 13.4%, 16.7%, and 16.4%, respectively. Worse yet for Dunn, who was always in the market (never in cash), was his firm’s “0 and 25″ fee structure. His trading revenue for those three years, in other words, was non-existent. (Note to hedge fund managers: Dunn’s internally managed confidence strategy has survived unscathed.)
Some traders would have closed up shop or perhaps faked their own death. But Dunn didn’t flinch. Despite three consecutive double-digit down years, he stuck to his reversal-based trend-following system — which, in the pursuit of lasting market moves, often sends him from net long to net short in one fell swoop, and vice-versa. His average return since 1974 is around 15%.
Dunn’s steadfastness led to a reversal of fortune. In 2006, he notched an up year. It was nothing spectacular — 3.1% — but he had stopped the bleeding. Then, in 2007, he produced a return of 7.6%. But true vindication has come this year: Dunn, making all the right moves as commodity prices have surged, has pounded out a 37.4% return through the end of June.
His year-to-date performance ranks his Stuart, Florida-based operation fifth among CTAs with more than $50 million, according to Institutional Advisory Services Group. By following his original system, just as he has been doing for the last 34 years, Dunn proved unflappable in the face of turmoil.
“The best I can do each day is follow my rules,” he says. “So if it didn’t work out today, big deal. There was no guarantee that it would work anyway. We never override the system.”
Not one discretionary trade since 1974? (Well, there was one during Y2K.) That’s practically unheard of. Is Dunn never gripped by a desire to take a flyer once in a while?
“No,” he insists. “What’s the point? Why do people think they are smarter than the market long term? What gives them that confidence? I guess people feel dumb if they can’t predict what the market is going to do in the short-term. They’re too proud to admit they don’t know what to do when they’re wrong. They don’t have the capacity to understand the digits that are scrolling by on the bottom of the television. I don’t. It’s too much noise. That’s why we rely on our system.”
Dunn’s approach to the markets has always relied on logic, order, and numbers, leaving bravado and emotion for the suckers. Born in Alton, Illinois, he was an undergrad University of Kansas (where he studied engineering physics) and obtained his PhD in theoretical physics from Northwestern in 1966. He taught physics at the University of California and Pomona College and then did top-secret operations research and the Department of Defense. “I found the work entirely tedious,” he recalls.
In 1970, he discovered something much more intellectually stimulating: the US stock market. But there were simply too many individual stocks to follow with any meaningful focus. He had one other problem: highly intelligent and dead set on making sense of the market, he couldn’t find a broker or money manager who impress them enough to enlist. “I had no confidence in any of them,” he says.
He read a newsletter — he can’t remember it’s name — about making money and commodities, which gave him an interest in technical analysis. He set out to develop his own system that would apply to all commodities. “There were only two dozen futures traders doing that back then,” he says. “I could keep track of the prices in my head, for heavens sake. And guess what? It worked.”
Today Dunn’s model tracks and trades more than 50 commodity and financial futures contracts. He has programmed his system to be statistically robust; his signal generator is constructed with only two parameters, which he refuses to reveal. He will, at times, adjust positions as volatility changes. For example, once a target risk profile is set, the system might buy 400 gold contracts when he puts a position on, then trim it by 10 or 20 contracts as volatility increases to better control the overall risk to the portfolio. He employs a macro overlay that mimics a value at risk (VaR) at the portfolio level. At that point, “we try to stay that course.”
Over the years Dunn has tweaked the model, but for the most part he has stuck to the game plan. He has a material amount of his own net worth committed to his system, and has never felt the need to sweet-talk clients, even during the dark days of 2003 to 2005. “We interview them as much as they seek information about us,” he says. “There’s no ambiguity — we tell clients that they have to be with us for the long haul, or we do not want their money.”
As a result, Dunn’s clients are loyal. Few bail out. “They leave because they want to buy a ranch in Colorado,” he says. “Or they die.”
“No one talks more straightforwardly than Bill Dunn,” says Ted Kingsbery, a principal at Dallas-based CTA Liberty Funds Group, who has known Dunn since 1981. “He’s an impressive guy. You know he believes in his system — he lives it.” Dunn, Kingsbery says, issues his statements on the first of the month (“the fastest I’ve ever seen”) and the figures, no matter how ugly, will be honest.”
As far as Dunn is concerned, his system doesn’t fail; it performs within its intended expectations — that is, Dunn knew, based on historical guidelines, that’s such an extended downturn was possible. But he doesn’t sweat the short-term; he doesn’t pour over tea leaves or act on cheap sentiment. So while the rest of us follow Obama/McCain and hurricane season, Dunn will merely be following his system. Just as he always has.Read More
In Broke: The New American Dream, you’ll see traders, poker players, investors, and politicians. Each has a unique manner in earning money, yet each relies on the concept known as mathematical expectation.
If you want to trade, learn this important concept first. Knowing it will explain the argument of Accuracy versus Expectation.
Accuracy Model: a high percentage of winning trades. Very hard to do for any period of time.
Expectation Model: a lower percentage of winning trades, 40% for example, but winners that are several multiples the size of the losers. Easier to accomplish in practice, although harder to deal with emotionally b/c our school systems have ingrained in us that 90% is an “A” Student, and 40% is a failure. Batting .400 though is Hall of Fame-level performance.
Mathematical expectation is used by everyone from the folks state lotteries, in all forms of advertising, Las Vegas and Atlantic City, insurance underwriters, and Wall Street traders. Without it, we would not know if our models would be profitable. It is an important tool in helping us put our statistical analysis to work.
Here is a brief lesson in how mathematical expectation works.
On the roulette wheel there are 36 numbers, double zero, and the blank. That makes 38 spaces to bet on. Each bet costs $1 to play. The winner pays $35. To calculate the mathematical expectation of the roulette wheel you do the following:
Multiply the probability of winning by what you win when you win. And from that, you subtract the probability of losing by the cost of each bet. The difference is the mathematical expectation. If it’s positive, it’s a fair bet. If it’s negative, you don’t play.
[(1/38) x (35)] – [(37/38) x (1)] = mathematical expectation of playing roulette.
(35/38) – (37/38) = (-2/38) or (-1/19).
So in the case of the Roulette wheel, the best bet is not to play. The problem is playing Roulette is fun! Most professional money handlers don’t find losing money fun. And that’s the difference between the professional and an amateur.
The roulette wheel, keno, and state lotteries are examples of games with negative expectation. Another famous one is the slot machine. If you go to Vegas with a budget constraint — and fixed amount that you’re willing to lose — and you play a game of negative expectation, you are guaranteed to go home broke. You can develop a slight edge in poker and black jack.
Professional traders develop models to buy and sell securities or commodities. They create entry rules, exit rules, and position sizing rules. They run 10 to 20 years worth of data through the model and come up with hypothetical results. Of course they use professional simulation software. The models can take several minutes to run.
The hypothetical results include the number of winning trades, the number of losing trades, percent winning trades, the percent losing trades, the biggest loss, the biggest win — just to name a few. With this data, traders can calculate the mathematical expectation of a trade and determine whether or not the system is worth following.Read More