Ed Seykota — Systematic Trend Following and the Cut-Your-Losses Rule
Ed Seykota · Early-1970s pioneer of computerized trend-following systems ~ present (Market Wizards 1989 / The Trading Tribe 2005)
Published 2026.07.08
Ed Seykota was the first to automate a trend-following system on a computer in the early 1970s. His method boils down not to flashy forecasts but to three disciplines: "react to the trend, cut first when you're wrong, and risk only 1%." This article is an educational breakdown of the structure of the method he publicly documented and explained in <b>Market Wizards</b> and Trading Tribe. It is a lagging observational framework, not a buy-or-sell instruction, and no rule guarantees a profit.
- Reaction, not prediction — you 'confirm' the trend direction with lagging signals like EMAs and Donchian channels, and only take positions in that direction.
- The core isn't the entry but cutting losses — a stop the moment you enter, less than 1% risk per trade, and speculative capital under 10% of net worth.
- Reward-to-risk, not win rate — an expected-value structure where a few big trend winners cover many small losses.
- The biggest risk is psychology, not the market — 'knowing' the rules and 'keeping' them are different things (Trading Tribe).
1. Identification — trend direction is 'confirmation,' not 'prediction'
Seykota's starting point isn't guessing direction. Using lagging indicators like the exponential moving average (EMA) or the Donchian channel, he reads only which way the trend is running 'now.' Long in an uptrend, short in a downtrend, and nothing when there's no direction. In his own words, he 'reacts to the price now' rather than divining the future.
Moving averages and breakouts are products of past prices. So the signal always arrives a little late, and in directionless stretches it fires false often. This is where the Seykota method both earns its 'slow but big' returns and gets its vulnerability in choppy markets.
2. Entry frame — react mechanically only to system signals
Entry is a rule, not discretion. You go in with the trend only when a pre-defined system signal — a new-high/new-low channel breakout, an EMA cross, and so on — appears. No counter-trend bets, and no averaging down to lower your cost on a falling asset. When there's no signal, standing aside is itself the position.
- On a higher timeframe, confirm that the trend direction is aligned one way per EMAs and channels
- Wait for the pre-defined signal in the direction of the trend (a channel breakout or cross) to be 'confirmed'
- Before anything else, fix in numbers the stop you'll set at entry and 1R (the entry-to-stop distance)
- If there's no signal or the direction is mixed, don't enter
What's laid out here is only the structure of the method Seykota publicly documented and explained. It is not a signal to buy or sell any particular stock or coin right now, and no adjustment guarantees future prices or profits.
3. Risk & sizing — nail down 'how much you'll lose' first
The heart of the Seykota method isn't the entry but cutting losses. He exposed less than 1% of the speculative account to risk on any single trade and capped the speculative capital itself at under 10% of liquid net worth. The moment he entered, he set a stop, and when the signal was wrong he got out without delay.
'The elements of good trading are cutting losses, cutting losses, and cutting losses.' The repetition is deliberate — the discipline of keeping losses small, more than entry technique, decides long-term survival.
Losing small when you lose comes before winning big when you win. The outcome is decided by a reward-to-risk ratio in which a few big trend winners cover many small losses — a matter of expected value, not individual hits.
4. Invalidation and limits — a stop hit is an 'error'
Touching the stop line means your read on the trend was wrong. You close without argument. When whipsaws (frequent stop-outs) pile up, cut size or stop trading altogether — Seykota went as far as to say, 'to avoid whipsaws, stop trading.'
1) Moving averages and breakouts are lagging indicators, so whipsaws multiply in choppy markets. 2) The widely cited '$5,000 to millions' record is a self-reported, model-account figure with no third-party audit, so it's hard to compare directly with regulated funds. 3) The discretionary override of 'knowing when it's okay to break the rules' takes years of experience, and beginners who imitate it end up with a collapse of discipline. 4) The psychological methodology (Trading Tribe / TTP) relies on participants' self-reports without a control group. 5) Most people, even knowing the rules, lack the psychological discipline to keep them and so don't reproduce the same results.
5. When you carry it over to crypto (perpetual futures)
The skeleton — align with the trend, cut losses, risk 1% — can carry over to crypto, and on the capital-preservation front it matters even more. But the market's nature is different, so you can't use it as-is.
- Timeframe: in a 24-hour, high-volatility market, lower-timeframe whipsaws are severe, so confirm the trend and signals on higher-timeframe closes such as the daily and weekly
- 1% risk: calculate off the 'actual loss amount,' not the leveraged notional value. Lower the leverage so the liquidation price doesn't fall within the stop distance
- Funding & liquidation: because funding fees eat away at the expected value of holding long, factor them into the reward-to-risk ratio, and keep in view the warning that averaging down at high leverage can turn into a 'total loss' via forced liquidation
- Signal filter: because stop hunts and false breakouts are frequent, cross-check breakouts against volume, funding, and open interest (OI)
Seykota was originally a systematic trader in the spot and futures markets and never issued a buy call on any individual coin. Borrow only the risk and trend frame; imitating 'concentration' through leverage sharply raises the risk of forced liquidation. This material is an educational breakdown of structure, not an entry instruction.
Let's back-calculate the reward-to-risk with hypothetical numbers (an expected-value view, not a win-rate one). With a $10,000 account, fix risk per trade at 1%, or $100. If the entry is 100 and the stop is 96, then 1R is the entry-to-stop distance, "4." Since the stop is 4% wide, the position's notional works out to $100 / 0.04 = $2,500 (25% of the account, unleveraged). You don't fix a target; as long as the trend is alive you trail it, leaving it open to 3R (target 112, profit $300) or 5R (target 120, profit $500). In a Seykota-style structure, even if losing trades outnumber winners, several -1R losses can be covered by a single +3-5R winner so the overall expected value comes out positive (+). The core isn't "how many times you're right" but "cutting at -1R when you're wrong and riding to how many R when you're right." The numbers are only an example, not a specific entry instruction.
- Is this spot a clear trend, or directionless chop? (If it's chop, standing aside is the answer.)
- Before entering, did you write down the stop (the invalidation point) and 1R in numbers?
- Does this trade's risk stay under 1% of the account? Does speculative capital stay under 10% of net worth?
- If the stop is hit, can you exit without emotion, or are you already thinking 'just a little more'?
- Have whipsaw losses been piling up lately? If so, shouldn't you cut size or stop?
FAQ
Does the Ed Seykota method have a high win rate?
It's not a method you judge by win rate. It's a reward-to-risk, expected-value structure that cuts losses short and covers many small losses with a few big trend gains. More than individual hits, the core is 'how small you lose when you're wrong,' so even over stretches with more losses, a few big winners can make the overall expected value positive (+).
Can you use it as-is in crypto (perpetual futures)?
The skeleton (align with the trend, cut losses, risk 1%) can carry over, but because of 24-hour trading, chop, funding, and leverage liquidation you have to re-tune the timeframe, signal filters, and sizing conservatively. It's a lagging observational frame, not a buy signal, and no adjustment guarantees future prices or profits.
Is Seykota's $5,000 to millions record credible?
It's widely cited, but as a self-reported, model-account figure with no third-party audit, it's hard to compare directly with regulated funds. It's safer to view the 'structure and discipline' for educational purposes rather than the performance itself.
What is the Trading Tribe (TTP)?
It's Seykota's psychology workshop and process that addresses the emotional problem of not keeping the rules even when you know them. It starts from the view that system failure comes more from a trader's 'departure from the rules' than from a flaw in the system. That said, since it relies on self-reports without a control group, treat it as reference only.