Erle Keefer: The Turtle Trained by Richard Dennis Who Questioned the Rules

Excerpts taken from The Complete TurtleTrader by Michael Covel:

Erle Keefer was the oldest Turtle hired. He was about the same age as Dennis, thirty-seven years old, when he started out as a Turtle. Describing himself physically he chuckled, “I’m a miniature Rich.” He also may have had the most diverse work experience prior to his Turtle time. He was a founding member of the London International Financial Futures Exchange, serving on the original membership and rules board. He received his undergraduate degree from the U.S. Air Force Academy and flew Air Rescue “Jolly Greens” in Vietnam. Military service was instrumental in his development. He said, “Combat forces you to be 100 percent ‘in the now.’ It changes your DNA forever.”

What Led Keefer to the Turtle Experiment

Keefer’s path to Dennis and the trading experiment was pure coincidence. He was sitting in a New York City sauna when he picked up a newspaper and spotted the Dennis ad. At that moment the female star of the movie Trading Places, Jamie Lee Curtis, was sitting in the same sauna with her boyfriend. Keefer was sitting there reading Barron’s. “I am looking at this ad and I knew who Rich was. I said, ‘Wow! This guy did it.’ ” Keefer thought there was little chance he would get accepted.

Keefer’s Desire for Proof

Erle Keefer, hired in the second Turtle class, was in the minority with his desire for a “proof of concept.” Like his mentor, Keefer remained skeptical: “You said I learned about the golden goose, but I didn’t have a computer program, hadn’t seen printouts, hadn’t seen proof of concept. You know what I mean?”

While Dennis and Eckhardt always knew what the Turtles were doing, it turned out that the two mentors didn’t always know what they were doing. During all that downtime of waiting for the markets to trend, four Turtles from the second class (Tom Shanks, Paul Rabar, Erle Keefer, and Jiri “George” Svoboda) formed a small group to do trading research. They wanted to validate their rules instead of just playing by them. While the rest of the Turtles may have been reading the sports pages and playing Ping-Pong, this group spent their time building a systems testing platform. Doing that took them a year. The results of their research project shook the program. They determined that Dennis had everyone taking far too much risk. The Turtles had all been trading according to rules taught by Dennis and Eckhardt and were making millions, but the research team, using original Apple computers, blended the S1 and S2 trading systems together and found that instead of a worst-case 50 percent drawdown, they were consistently getting a worst-case 80 percent drawdown.

What the Keefer Research Group Found

The finding that the combined S1/S2 system produced worst-case 80% drawdowns rather than the 50% Dennis described is the most operationally significant research result produced within the Turtle program. An 80% drawdown requires a 400% gain to recover. A 50% drawdown requires only a 100% gain. The difference between these two recovery requirements is not academic. A system that produces an 80% drawdown during an adverse market period is a system that most investors cannot sustain psychologically or financially, even if it is ultimately profitable over the full cycle.

Keefer’s background in the Air Force and at LIFFE gave him a different relationship to rules than most of the other Turtles. A pilot does not follow pre-flight checklists because he is told to. He follows them because he understands what each item checks and why it matters. The “proof of concept” demand reflects this orientation: before trusting your life, or your capital, to a set of rules, you want to know that the rules have been validated against real conditions, not just asserted by an authority figure.

The year-long system testing effort during the downtime while the markets were not trending is the model for how systematic trading research should be done. The four Turtles were not optimizing for historical performance. They were stress-testing the system to find its failure modes. They were asking: what is the worst this system can do, and can we sustain it? The answer they found, 80% worst-case drawdown, was worse than anyone had been told. That finding led to adjustments in how the system was applied that protected the Turtles’ trading capital during subsequent adverse periods.

The “100 percent in the now” formulation from Keefer’s military experience is also the complete description of systematic trend following’s operational requirement. Combat forces awareness of what is actually happening rather than what was expected to happen. Price-reactive systematic trading does the same: it responds to what price is actually doing rather than to what the analyst expected it to do. The military training that Keefer brought to the Turtle experiment was not incidental to his trading performance. It was the psychological foundation that allowed him to apply the rules consistently in the uncertainty of live market conditions.