StratoQuant
EN
Get the free code

F-03 · Foundations · system design

Breakout systems: buying the escape from a range

What a breakout system does, how its rules are built, why most breakouts fail, and what it takes to catch the few that run.

By Baris Demirci · · Rev00


The idea

Markets spend long stretches moving sideways inside a range. Sooner or later price escapes it, up or down. A breakout system waits for that escape and joins it, on the view that a move strong enough to break out of a range may be the start of something larger.

Breakouts are one of the oldest ways to enter a trend following system. Richard Donchian traded channel breakouts in the 1960s, and the Turtle traders of the 1980s made them famous. The same idea is also used over much shorter periods, down to the range of a single morning.

What the rules look like

A breakout system usually has four parts:

  1. A definition of the range. Most often the highest high and lowest low over a set number of days, sometimes called a channel. Other versions use the first part of the day’s trading, or a band around an average.
  2. A trigger. Price closing above the top of the range, or trading through it.
  3. An initial stop. Where the trade is wrong. Often the other side of the range, or a set multiple of a typical day’s move.
  4. An exit. A breakout the other way over a shorter period, a trailing exit, or a set number of days.

Many versions add a filter, for example taking breakouts only in the direction of the longer-term trend, or only after a quiet period when the range has narrowed.

Held Recent high Recent low Initial stop Buy on the break Sell if price falls back to the stop
Illustration, not market data. Price swings inside a range, then closes above the recent high. The system buys there, places its initial stop at the bottom of the range, and holds while the move runs.

Most breakouts fail

Price pokes above the range, the system buys, and price falls straight back in. That is a false breakout, and it is the normal case, not the exception.

Held Recent high Stopped out, small loss The next break
Illustration, not market data. A false breakout: the system buys the break, price falls back into the range, and the stop takes a small loss. The next break may be the one that runs, and the system has to be there to take it.

So breakout systems have the same shape of results as other trend systems: many small losses from failed breaks, and a few large gains from the breaks that keep going. The whole record rests on taking every break the rules give, because nobody knows in advance which break will be the one that runs. Skipping the next one after three failures is how most people miss it.

On the close, or during the day?

There are two ways to act on a breakout, and they behave differently:

  • During the day. A stop order sits just above the range and fills the moment price trades through it. It catches the move early, but it fills on brief pokes that reverse within the hour, and it often fills at a worse price than the level, especially on a fast move.
  • On the close. The rule checks the closing price, and the order goes in at the close or the next open. It ignores breaks that do not hold to the end of the day, but it pays a later price. Acting at the close usually means a market on close order, and acting at the next open a market on open one.

Either way, a market that gaps through the level fills at the gap, not at the level. A backtest that assumes every breakout fills exactly at the top of the range will look better than the system can ever be.

A breakout system, step by step

After the close: update the levels Holding a position? Close above the recent high? Close below the exit level? Buy next open, set the stop Do nothing Sell next open Raise the exit, hold No Yes Yes No Yes No
Illustration of a typical daily routine for a long-only breakout system that decides on the close and acts at the next open. Every branch ends in a definite action, including "do nothing". Real systems differ in the details.

What it asks of you

  • Buying at a high. A breakout buys when price is higher than it has been for weeks. It feels expensive every time.
  • Taking every break. Most fail, and the system only works if you are there for the few that do not.
  • Resisting filters. After a run of false breaks, it is tempting to add a filter that would have avoided them. Each filter fitted to past failures spends a little of the evidence, and enough of them leave a system that only works on the past.
  • Paying for fills. Breakouts buy into moving markets, so slippage is usually worse than the backtest assumes. Test with realistic costs.

The short version

A breakout system waits for price to escape a range and joins the move, with a stop where the idea turns out wrong. Most breaks fail for a small loss, and a few run far enough to pay for all of them. It works only if you take every break, test with honest fills, and resist the urge to filter away the last few losses.


Written for people building their own systematic portfolio management process. Nothing here is a recommendation to buy or sell anything. See the disclaimer.

← All articles