Trading systems

Section 01

Trend following

You enter in the direction of a move that is already underway, and you try to stay in until the move is clearly over. Entries are often a break beyond a channel, or a cross of a slower average. The classic pain is a long row of small losses while markets chop, paid for by a smaller number of large wins. If you need to be right often before you can sit still, this style will push you to quit right before the outlier that pays for the rest.

A famous historical case is the turtle experiment. Richard Dennis and William Eckhardt trained a group of novices in the early 1980s to trade futures with written rules. Those rules included breakouts of Donchian channels, the highest high and lowest low of a lookback such as 20 or 55 days, and position size based on N, a volatility measure in the family of average true range. The episode is evidence that a simple rule can be taught. It is not evidence that those lengths pay the same way in your market this year. Trend-following programs have endured long, public drawdowns since. The style did not die. The comfort never existed.

Section 02

Mean reversion

You bet that a stretch away from a recent average will relax back toward it. The average might be a moving average, a day's volume-weighted price, or the middle of a range. Win rates are often higher than in trend following, and the wins are often smaller. The catastrophic version is averaging down into a trend because each new price "is even more oversold." Oversold is a description of the past window. A market can stay stretched until your account cannot.

A mean-reversion rule needs a hard invalidation more than a trend rule does, because the comfortable action, adding, is the dangerous one. Decide the maximum loss before the first entry. Include a rule for news that can reprice the instrument in one print. Fading a quiet range is a different trade from fading an earnings gap.

Section 03

Breakout

A breakout is a bet that a push out of a defined area will continue. The area might be the day's opening range, a prior day's high, or a pattern boundary from the chart-pattern page. Toby Crabel's work made the opening-range break a named template: define the range of the first minutes, then trade a move beyond it. The template is not a gift. The cost of breakouts is the false break, and on some days that cost is most of the trades.

Write whether you need a close outside the area or a tick outside it. Write whether volume must expand, and write what you will do in a market where you do not trust the volume figure. Then write the snap-back rule. Without that third sentence, a breakout system is a hope that momentum continues, expressed with money.

Section 04

Pullback in the direction of the trend

You mark a trend on a higher time frame. You wait for a retracement into an area you defined, such as a prior swing, a short average, or the session's volume-weighted price. You enter only in the direction of the higher-time-frame trend. You will miss the moves that never pull back. Missing them is the point. You are paying that opportunity cost in exchange for not buying the extended print.

The failure mode is a pullback that is actually a reversal. Your stop belongs where the higher-time-frame idea, not your pride, is wrong. If you keep widening it "because the trend is strong," you have retired the system and kept the slogan.

Section 05

Relative value, briefly

A pairs or spread trade is long one instrument and short a related one, betting on the relationship rather than the direction of the whole market. Correlation of the two returns can be high while the spread still trends against you for longer than you can fund. Long-Term Capital Management is the cautionary study on the mastery page: leverage turned a convergence idea into a survival problem. Do not start here. If you ever do, the position is two risks, two borrows, and one hypothesis, and the size should assume the relationship can get worse before it gets better.

Section 06

How to choose

• Time. Trend following on daily bars can be run after work. A one-minute opening-range rule cannot.
• Temperament. Count how angry a string of eight small losses would make you. If the answer is "I would double the next one," you are not a breakout trader yet. You are a person who needs a smaller size or a different style.
• Costs. Mean reversion that aims for a small move dies in a wide spread. Trend following that holds for weeks cares more about the big error than the penny.
• One style. Run one written system until the journal has something to say. A second system is allowed later, as a separate set of rules, not as a way to take the other side of your own trade because you are bored.