Breakout forex strategies – what you need to know

Hello everyone!

Many beginners often think that a trading strategy should work always and under any market conditions. Also, some have little idea what types of trading strategies exist and what are the peculiarities of trading for this or that type of TS. Knowing the peculiarities of each type of strategies, their advantages and disadvantages will allow you to develop more thoughtful and high-quality TS’s. And today we are going to discuss the peculiarities of building breakout forex systems.

The breakdown of trend lines and important price levels means the appearance of many large players in the market with clear preferences for the growth or decline of this or that instrument. It is at such moments that signals are generated, helping to enter the market in the direction of large volumes. These are, as a rule, classic trading strategies, tested by decades of trading on various instruments.

The main idea behind breakout trading strategies is that if the price is in a range for a long time, there is bound to be a moment of forceful market decision and a breakout in a certain direction. Therefore, for this type of strategies it is quite critical to choose the very levels for breakout and correct filtering of false breakouts.

In order to fix a true breakout, we need to correctly realize the direction of the current trend, correctly select levels, the breakout of which can lead to a serious price change and timely determine the truth of the breakout of these levels. That is, we need less noise that can distort our analysis and contribute to a wrong trading decision. That is why breakout strategies work best on instruments with volatility not lower than average and on periods not lower than hourly, where there is not so much noise.

I do not say how true it will be, but let’s agree within the framework of this article to call a system type a set of different systems identical in their basic idea: trend, reversal, breakout. That is, the type of trading system answers the question “What are we going to trade?”. We will call the system class a set of trading systems of the same type, which use a common approach to finding trading signals. The question “How will we trade?” is more appropriate here. So, what classes can there be within such type of systems as breakout systems?

This is one of the oldest breakout trading patterns. On a rising trend, a line is drawn on rising lows, and on a falling trend, a line is drawn on falling highs. As for all classes of breakout strategies, it works well on the period from H1 and above.

Of course, the market is not always in a trend and it is not always possible to find a suitable trend line. And even if a directional movement is found, trendlines are rather subjective. Therefore, this class of trading systems is almost impossible to test – if you repeat the test several times in a program for manual testing, for example, Forex Tester 3, you are likely to get slightly different results.

There are also quite a few indicators that detect trendlines automatically, but I haven’t come across any that do it well.

This class of breakout systems is also a classic and one of the oldest. The signal to enter this strategy should be not only the fact of breaking a certain level or range, but also the analysis of the market behavior near the consolidation boundaries. In such strategies, stops are set, as a rule, at the last level of the maximum or minimum before the breakdown, and profit is taken in the amount of the amplitude of the consolidation range or the previous movement.

There are always support/resistance levels, on any asset and on all timeframes. A true breakout of a level means a strong trading impulse, which gives the opportunity to take profits with minimal risk.

Fibonacci levels can also serve as excellent entry points for a breakout. I did not separate such systems into a separate class, as they are also just price levels.

As a rule of thumb, wait for the breakout of Fibonacci levels such as 38.2 and 50. Such a strategy will not necessarily be counter-trend, because the levels can be built on the basis of pullbacks in the existing trend. The ZigZag indicator is usually used to determine the basis for building levels. It helps to perform all the constructions automatically, thus avoiding subjectivity.

There are many tools for building various levels. On the forum in the indicators section you can find pivot level indicators of different ways of building: round levels, Murray levels, Fibonacci levels, levels built by Fractals, and others – there are hundreds of them. And this gives a great scope for researching the effectiveness of using this or that way of defining levels in order to find the most effective way of their construction for such a class of strategies. Perhaps, there lies an indicator that is ideal for building an Expert Advisor based on the system of breakdown of support/resistance levels.

The SMA and EMA lines, especially the standard set with periods of 20, 50, 100, 200 are the strongest price levels, and their breakdown means the exit from the average value zone and the formation of a new trend.

As a rule, the signal is generated when closing above the maximum or below the minimum of the breakout candle. Nevertheless, this class of systems in its basic version currently brings quite a lot of false signals – the markets have changed a lot with the advent of such a phenomenon as online trading and with the ability to perform complex mathematical calculations literally instantly. Therefore, this class of systems, like the other classes, has started to evolve. But unlike the other classes, it has evolved into a completely new class of systems – volatility breakout, which we will talk about a little later.

Despite the fact that this approach is considered outdated, there are a huge number of different moving average indicators and even if you cannot build a profitable strategy of this class, its development will give you a lot of useful experience. On the forum there is a set, also compiled by Pavel, with more than 600 different moving average indicators for every taste.

In fact, there is no need to use moving averages. You can work with an indicator such as Envelopes, taking trades to buy when the upper boundary of the indicator is broken and vice versa for selling. In essence, it will remain the same class of trading systems, just slightly modified.

Historically, trendline breakout patterns have been followed by channel breakout patterns, which are based on support and resistance lines calculated from past highs and lows. A trader buys when prices rise above the high of the last n bars (the upper boundary of the channel) and sells when prices fall below the low of the last n bars (the lower boundary of the channel). Channel breakout systems are easy to program and very simple to understand. The simplest and most common example of a channel is the Donchian channel:

This channel is built on the highs and lows of a certain number of bars, in our example 60 last bars. When this channel is broken, the market is entered.

The most famous example of this type of strategy is Richard Dennis’ Turtle System. Like all breakout strategies of this class, it is very simple in terms of automation, as you have seen when writing an Expert Advisor using this system. Nevertheless, the basic system shows good results up to 90-95 of the last century and without various modifications in the last twenty years practically does not show profit, forcing traders to complicate and improve their TS.

As a rule, the exits for such systems are realized at the breakdown in the direction opposite to the open position. In the Turtles system, entries are made at the breakdown of the 60-day Donchian channel, and exits at the reverse breakdown of the 20-day channel. Protective stop loss orders are best set with reference to volatility, for example, according to the ATR indicator. The best results are shown by the coefficient within 2 – 3 ATR, depending on the instrument. Alternative exiting of positions with a binding to the time of holding a deal is also often used. For example, all positions that were opened, say, 30 days ago are closed. This approach helps when the system gets into a sideways movement and hangs around for a long time without a significant increase in profitability.

This class of trading systems is actually profitable if all the elements of the system are thoroughly worked out. The basic concept, as I said, ended the period of good profitability in the last century.

Another interesting variant of building a price channel is by session or by range within certain hours. The BigDog strategy is an example of such an approach. According to the rules of the strategy, a price channel is defined with a starting point 3 hours before the opening of the London session and an end point at the opening of the session. Thus, we get the price channel of the morning calm, on the breakdown of which we will trade.

The trading system does not show much high efficiency, but, if properly finalized, it can serve as a basis for strategy development. To work on it, it is recommended to use short trailing – within 10-15 points. It is also advisable to try to take trades in channels not exceeding 20 pips in height and avoid trading in channels over 50 pips.

Another option is to use the highs and lows formed, for example, during the Asian session. This approach is called a breakout of the highs/minima of trading sessions. The breakdown of the highs and lows of the Asian session can be considered the most profitable.

Another interesting variant of building a price channel is using the Ishimoku cloud:

Of course, it is not a grail, nevertheless, the approach is original, interesting, gives decent accuracy, good profitability and risk to profit ratio. If carefully developed, a trading system using this idea will be able to enter trades at the very beginning of new trends and exit actually at their completion. I think that very few big players use this methodology, so it remains effective to this day. By the way, the author of this system is a famous Swedish trader and analyst Lars Larsson, who claims to have achieved 80% of profitable trades in binary options trading with this modified strategy.

The modified version uses filtering of trades by the Awesome Oscillator indicator, but for binary options the number of profitable trades is critical, not the ratio of profit to loss. Therefore, for Forex, you can do without additional filtering, having carefully considered the support of trades, possible profit targets and loss limits. To be fair, it should be noted that the author offers several different options for entering a position – the breakout variant, which we have analyzed, and the trend variant, which is based on the reverse crossing of the lines forming the cloud boundaries.

The class of systems based on the breakdown of price channels works better than all other classes of breakout strategies in the tests, despite the sometimes not very stable results.

Newer and more complex are the volatility breakout models, where the points whose crossing triggers a signal are based on volatility boundaries. Volatility boundaries are located at some distance from the current price (for example, the closing price), and the distance is determined by the current market volatility: when it rises, the boundaries move further away from the current price, when it falls, the boundaries narrow.

Most often for measuring volatility in this class of systems the ATR or Std Dev indicator is used. Thus, the Bollinger Bands indicator is built to calculate the channel boundaries by standard deviation, and the Keltner Channel indicator is built on ATR.

It is based on the following statistical idea: if the market moves in a given direction stronger than expected from a normal oscillating movement (which is reflected in the volatility), then, perhaps, there is an influence of some force, i.e. a real trend. Buy – when the price rises above the volatility limit, sell – when it falls beyond the lower limit.

Many of the systems of this class were very popular in the late 80’s, but nowadays they are quite rare. Personally, I have never been able to get this class of systems to work satisfactorily, but maybe I just can’t design effective filters to minimize false breakdowns. Or perhaps indeed the current market for such systems is not suitable. This assumption is also supported by many market researchers and their independent tests of the class of such systems for the currency market.

This class of strategies is based on the breakdown of levels in the readings of one or another oscillator. An example of such a strategy is the Morning Stars Rustle for the H4 period, which is based on the breakdown of the levels of Bill Williams’ Awesome Oscillator indicator.

Another option is the breakdown of overbought/oversold levels:

This example uses the breakout of the 100 and -100 levels of the CCI indicator. A moving average is used to determine the trend and filter trades against it. For this class of strategies it is not very important which oscillator to choose, it is much more important to maximize the effectiveness of the method for determining the direction of the current trend.

The idea here is exactly the same as for all other breakout trading strategies – if the price has broken through an overbought or oversold level, it is very likely that this movement will continue.

To exit a position, as a rule, it is recommended to focus on the reverse crossing of the previously broken level by the oscillator.

Breakout-based models can also differ in the method of market entry. Entry can take place at the opening or closing of the day, or intraday entry using orders at boundary levels. More complex methods allow you to buy or sell on the boundary, i.e. to try to enter the market on a pullback, when after breaking some boundary, prices briefly return to it.

The simplest variant – entering on the opening of a new candle, allows you to test breakout systems quite quickly and accurately, because for such variants are not important movements occurring within the candles of the selected timeframe.

When working with stop or limit orders, it is very important to use tick data and testers that support changes in the spread during testing, because inaccuracy of just a couple of tenths of a point can lead to the activation of a pending order or, on the contrary, to the skipping of certain deals. It does not matter what period the trading system is developed for – it will still be very demanding to the quality of testing and the results of trading at different brokers can be radically different.

That is why I personally do the following:

That is, after the candles have broken through a certain level, the strategy involves, for example, setting a pending stop order slightly below the shadow of the breakdown candle. Instead of placing an order, I simply start monitoring the level of the pending stop and as soon as the candle closes below it, I enter the trade on the market at the opening of the new candle. This saves me from having to use tick data and generally increases the reliability of the system, albeit at the price of a slightly delayed entry.

There is already a blog article on breakouts and how to enter trades. The classic definition of a breakout is simple: it is the crossing of a strong price level, after which the price is more likely to move in the direction of the breakout than to move back. For a breakdown to occur, something major must happen in the market to move the price some distance.

It turns out that if the price is approaching power levels, there are four possible behavioral scenarios:

It is believed that the more unsuccessful breakout attempts, including false breakouts, the more stable and significant a particular price level is considered to be, and the more market effort will be required to actually overcome it. A reliable breakout requires at least 2-3 tests on the period from H1 and above, and the more visible pullbacks, the stronger impulse should be expected in the direction of a new movement.

The following properties are attributed to a true breakout:

A false breakout is a price breakout of some key level, followed by a quick (1-2 candles) reversal in the opposite direction. The concept of a false breakdown is directly related to market psychology. It is a manifestation of the “herd” reflex in the market, when small players try to catch the outgoing movement without serious analysis. As a result, they buy at the tops and sell at the lows.

Here are some criteria for filtering false breakouts:

One of the problems with using breakouts is that there is a tendency for extremely “sawtooth” trading when the system registers a breakout and no real trend follows. One possible solution is to use a trend indicator to filter out breakout signals. Many traders use the popular ADX indicator.

Filtering out non-trending markets, “sawtooth” trading, and protracted trades will slightly improve the results of the system. ADX is used to filter out breakouts according to White’s research (White, 1993). A trend is considered to exist when the ADX, calculated from the last 18 days, reaches a new six-day high. Entries are made only when a trend is present.

But is ADX filtering so useful? We have already found this out in the article linked above. I would only add that for breakout systems the use of ADX can still slightly improve the final result.

The disadvantages of breakout systems include the fact that breakout models, as well as many other models based on trend following, can enter the market with a delay, and sometimes so late that the movement has already ended. In addition, a position may be opened on a small price movement that does not allow for profit taking.

Another hypothetical problem, according to many traders, is that with the widespread availability of high performance computers, simple breakout-based methods no longer work well enough. It is believed that due to the widespread popularity of breakout-based systems, markets have adapted to them, and the efficiency of using breakout-based TSs has decreased due to the increase in price noise near the boundaries where orders of breakout-based systems are placed. This circumstance makes these systems trigger too often, especially in active markets characterized by high volatility.

The general property of breakout systems is that they work better on long periods and trending markets. A properly constructed breakout trading system should address the problem of noise near market entry points.

A separate problem is the width of the channel for a breakout strategy. If the thresholds are set too close to current prices, there will be a large number of false signals, which will lead to sawtooth trading – price noise will trigger orders in one direction or the other. Since such movements do not represent real trends, the trades will at best not be profitable, hitting the trader’s capital.

If the boundaries are spaced too widely, the system will make too few trades and enter the market too late at any important movement. If the channel boundaries are set optimally, a theoretical system based on breakout patterns can be very effective: frequent and small losses caused by the lack of continuation of the movement or price noise will be compensated by significant profits at large market movements.

To reduce the number of false signals and reduce the sawtooth nature of trading, breakout-based systems are sometimes combined with indicators that presumably determine the presence or absence of a trend in the market. If there is no trend, the entry signals generated by the system are ignored; if there is a trend, they are accepted for execution. The problem is that indicators do not always function accurately or do not react fast enough, lagging behind the market and making the system not perfect. Otherwise, any trader who used them in combination with a breakout or other trend following model would get rich very quickly, as the trading strategy would only enter significant trends, trading smoothly and steadily.

Overall, simple breakout systems follow a preordained pattern and do not work well enough in today’s high-performance markets. This is consistent with the assumption that profitable trends are becoming fewer and fewer. Many traders believe that markets are becoming more “noisy” and counter-trending, making it difficult for the above methods to work. Exceptions are young instruments, like the same cryptocurrencies, as well as low-liquidity instruments such as USDMXN or USDZAR pairs.

There are many classes of breakout-based systems, many trend filters besides ADX and many additional ways to improve the proposed basic classes of systems that were not considered here. I hope I still managed to give you a good overview of the popular classes of breakout based trading strategies and a solid foundation for your own research.

Good luck and see you all again!