Ed Seykota built one of the first computerized trend-following systems for client money. His reported record is extraordinary, but it rests on his own account, not audited statements.
In the early 1970s, a brokerage firm put Ed Seykota’s computerized system in front of its clients. On paper, the system was heading for a gain of about 60 percent for the year. Then management interfered. By Jack Schwager’s account in Market Wizards, the year’s most profitable trade ended as a loss, and quite a few accounts lost money instead.
The episode explains the Ed Seykota trading strategy better than any single trade does. Seykota says his boss could not follow the system. A superior wanted it changed to generate more commissions. Seykota says he told them their best move was to make money for customers. Then he left, taking about six accounts of $10,000 to $25,000.
Those accounts later produced one of the most quoted numbers in trading. So how did he do it, and how much of the story can be checked?
The Losing Trade That Started It All
Seykota’s first lesson came before the computers. In the late 1960s he decided silver had to rise once the U.S. Treasury stopped selling it. He opened a margin account. His broker talked him into a copper short, and he was stopped out. Then he bought silver. The price fell, and his stop was hit.
He later called it a stunning education in how markets discount news. It pushed him toward price rather than opinion.
He graduated from MIT in 1969, having studied servo theory, the engineering of self-correcting systems. Around then he read Richard Donchian’s work on a diversified moving-average crossover system. An automatic money-making machine fascinated him. So he rented computer time, punched cards and tried to reproduce Donchian’s results.
At the wire house he ran weekend tests on an IBM 360/65 mainframe. He tested four system types, about 50 parameter sets and eight commodities across roughly a decade. The work took about six months. Other parameter sets also worked. Longer-term smoothing held up better, while transaction costs chopped up shorter-term systems.
How Ed Seykota Actually Made Money
Seykota made money by riding long-term price trends in futures and cutting losses before they grew. He did not try to predict prices.
The core was mechanical. He built on Donchian’s idea, tested variations and favored slower trends because costs punished short-term trading. His exact entry and exit parameters were never published. One source says his first system used exponential moving averages, but his own interviews do not confirm it.
Exits mattered most. He set stops before entering, at the point where the chart soured. As a trend extended, he moved stops closer to lock in profit. Losers stayed small and winners stayed open-ended, so a few big trends could pay for many small losses.
Judgment never left the process. Seykota said system traders still decide how to roll contracts, how much to bet and which markets to trade. Futures margin supplied the leverage, though no ratio is documented. His own theory of trends was behavioral. Prices keep moving while society gradually works out what a story means. When everyone understands it, the move ends.
Ed Seykota’s Trading Rules and Risk Management
Seykota’s rules center on small, predefined risk. The wording below comes from his interviews and a 1993 paper he co-wrote with Dave Druz. Exact quantitative parameters were not published.
| Principle | Market Logic | Modern Relevance |
|---|---|---|
| Set the exit before entry | A trend trade is wrong when the chart breaks. Price, not opinion, decides. | Timeless. Still the core of disciplined trading. |
| Risk under 5% of equity per trade | Gaps and poor fills can exceed a stop. The cap leaves room for losing streaks. | Timeless. It is a ceiling; his typical size is unknown. |
| Ride winners | Trends can run longer than expected, and they fund the losers. | Adaptable. Needs a tested exit method. |
| Cut size in drawdowns | Losing streaks are normal for trend following. Chasing losses compounds them. | Adaptable. No reduction schedule is documented. |
| Size by volatility | Each market gets similar risk. No single market dominates. | Adaptable. Now a standard tool. |
| Judgment sets the risk | A system cannot decide how much risk to accept. | Timeless. |
The volatility rule is the only one with a published formula. Bet size equals equity times a “heat” setting, divided by the number of instruments. Contracts equal bet size divided by the risk to a three-week price point. The paper also warned that its profits largely reflected trending commodities chosen in hindsight, and that results faded over time. It was not a performance claim.
Holding Through the Limit Moves
Seykota’s best-documented trades come from a friend, not from himself. Michael Marcus, who credits Seykota as a major teacher, described them in Market Wizards. These are his recollections. No sizes, prices or profits survive.
In the 1973 soybean boom, Marcus says he sold early. Seykota stayed long, and soybeans rose the daily limit twelve days in a row. Marcus recalls dreading each morning.
Then the reverse. Marcus says Seykota recommended coffee. Marcus bought with a close stop and was quickly stopped out. Seykota held no stop. The market locked limit-down for several days, and he could not exit.
The two stories show the method’s two faces. Staying in made the big win possible and the big loss unavoidable. The coffee story also sits awkwardly with his later statement that he sets stops at entry. The eras differ, and the exact orders are unknown. Marcus later said that watching Seykota stuck in coffee felt like joy, which he took as a sign he was too competitive.
What the Record Actually Shows
Seykota reported an extraordinary result in one client account. No audited statement, regulatory filing or fund document confirms it.
| Metric | Result | Context |
|---|---|---|
| Model account gain | Above +250,000% | One client account opened with $5,000 in 1972, measured to mid-1988. Chosen by Seykota. |
| Dollar profit | “Over $15 million” | His description. It does not reconcile neatly with the percentage, and the basis is not stated. |
| Duration | 16 years | The book’s preface says sixteen. Many sites repeat twelve with no source. |
| 1990s average | Nearly 60% a year after fees | Relayed from a 2001 Futures article that could not be retrieved. One version applies it to his whole career. |
| Drawdown, Sharpe, win rate | Not publicly verified | A “12% drawdown” on his website belongs to a reader, not Seykota. |
Schwager published the model-account claim in 1989, and Seykota’s own 1992 interview confirms the research story, the mainframe work and the dispute with his employer. His 1993 sizing paper is on record. Marcus, in print, credits him as a mentor.
The account’s statements, the withdrawals and the exact return basis. The 1990s figure. His assets under management. Any net-worth figure. Online claims of billions have no source.
He told Schwager that he does not publicize a track record beyond the model account. That choice makes independent checking impossible.
Students and the Trading Tribe
Seykota’s influence runs through people more than through published rules. Marcus credits him with teaching loss-cutting and trend patience. David Druz apprenticed with him in 1991 and 1992. In 1992 Seykota also began the Trading Tribe, a peer group for examining the emotions behind trading decisions.
Where Seykota’s Approach Could Break
Trend following struggles when prices chop sideways, and Seykota knew it. He noted that costs wear down short-term systems. His 1993 paper admitted that its results faded. The early-1970s override showed a second weakness: the method fails when people tamper with it. The coffee episode showed a third. A position can be impossible to exit.
In 1988 Marcus argued that trend following was finished. He cited crowded professional money, false breakouts and central banks leaning against currency moves. That was one trader’s opinion, and the model account was still reported up that year. But the doubt is documented.
Today’s markets add pressure Seykota never discussed. Many managers now use similar trend rules, so entries and exits cluster. Sudden reversals can hit several diversified markets at once. Simple moving-average rules are also widely known. These are modern judgments, not documented failures.
Can Ed Seykota’s Trading Strategy Still Work Today?
His risk principles still work. His exact system cannot be tested, because it was never published. The broad idea of trend following survives, but the easy edge in plain moving averages has probably shrunk.
| Principle | Still Relevant? | Modern Difficulty |
|---|---|---|
| Predefined exits and small risk | Yes | Discipline, not access. |
| Volatility-based sizing | Yes, adapted | Now routine, so it is not an edge. |
| Plain moving-average trend rules | Weaker | Widely known and heavily traded. |
| Early computer testing | No | Anyone can do it now. |
| Choosing clients who stay put | Hard | Retail traders cannot screen themselves. |
TraderVerified Verdict
Seykota’s real contribution was proving, early, that a tested set of trend rules could run client money. The opening story shows why it mattered. The rules worked only while people left them alone.
His reputation survives in outline: early system builder, strict risk discipline, influential mentor. The headline return does not survive verification. It comes from one account he chose, with no statements, and its duration and dollar figures conflict.
Modern traders should keep the exit-first rule, small risk and drawdown discipline. They should not copy a system nobody published, or assume a number nobody audited.



