How to trade pullbacks

I’m starting a short series of posts here on pattern recognition, focusing on simple patterns that I have found useful in actual trading. Some of my readers may have questions as to whether it is possible to trade profitably based on simple patterns, given that so much modern quantitative trading focuses on volatility, spread relationships between two or more markets, and other types of trading that move us away from simple, directional trading. The answer to that is a qualified ‘yes’: patterns are heuristics, shortcuts, if you will. A trader who can read and analyze patterns can quickly drill down to the essential elements of market structure. Opportunity exists in markets when buying and selling pressure is not balanced, and patterns can help us understand this critical balance very quickly.

Though many of these patterns have a slight edge on their own (read this, for pullbacks), I think far better results come with experience and when the trader learns to read the action in the market; we shift from a “take every pullback” to a “find the best pullback” mindset. Traditional technical analysis often focuses on visual symmetry, how the pattern looks on a chart, to find the best trades, and I would suggest that this might not be right. The best patterns are often pretty ugly, but they fall in beautiful spots in the market structure.

How to trade pullbacks

The key to trading pullbacks is that you are trading the fluctuations in a trending market. For this to work, it’s obvious that the market must be both trending and fluctuating. It is normal for trending markets to show pullbacks, but some very strong trends do not. Don’t try to force trades in this environment–you don’t have to be in any move or any trend; wait for the best trades. What might not be so obvious, at first, is that trading a pullback is a prediction that the market will continue to trend. I might suggest that this is slightly different from identifying a trending market.

From a simple perspective, trading pullbacks boil down to.

  • Wait for a trending market to make a strong move.
  • Look to enter the market when it comes back to some sort of “average” price.

Digging deeper

Both of these elements require a little more information, to be done well. First, you have to be able to read market structure and understand trends. It’s not enough to say “this market starts at the lower left of the chart and ends at the upper right”. That is enough to be able to identify high level trends, but, remember, you are looking for a market that will likely continue to trend. Being able to juggle elements like length of swing, lower timeframe momentum, exhaustion/strength will help, but, in my experience teaching and training traders, these are skills that develop naturally from trading pullbacks.

Identifying a market that has made a strong move can be done with the help of reasonably calibrated bands. There is a wide range of parameters and settings, but I’d suggest that whatever bands you use should contain roughly 80% – 90% of the price action. Consider Bollinger bands: if you set them 0.5 standard deviations, you are probably too close, and 5.0 standard deviations is probably too far. The first case will be hit constantly on trivial movements, while the last will only be hit with the most extreme exhaustion moves, but, in the middle, there is a wide range of values that can be useful. (I have found Keltner channels more useful than Bollingers, but I also think this is a matter of personal taste.) Play with moving average length and band width, settle on settings you will use, and don’t change them. The bands provide structure that will be critical as you develop your intuitive sense of market structure.

Note points where market touches bands (ES 11/4/14)
Note points where market touches bands (ES 11/4/14)

The chart above shows points where 2 minute (I focus a lot on higher timeframes, so let’s look at intraday data a bit, too) bars touch the Keltner channels. Consider this as setting up a trigger condition, and then we look to enter somewhere “around a middle.” Now, I’ve written at great length about how moving averages don’t work (here, here, and especially here), so am I telling you to buy and sell at a moving average? No, the average price is not important, but the concept of trading near the average is what matters. If you test this, you will find an edge whether you buy in front of or through the average, and you will also find an edge with many different lengths of moving averages. There is no magic to the average, but perhaps there is magic in a disciplined, consistent approach to trade entry. Look at the following chart, which shows points where we might have executed (long or short) near the average following the setup condition in the first chart.

Possible pullback entries
Possible pullback entries

Further refinements for entries

Without going into too much detail, it is worth considering some other elements that are important. The most important is to remember that you are looking for a market that you believe will continue to trend, for at least one more trend leg, so it makes sense to avoid patterns that point to end of trends. The most important of these is exhaustion. Avoid buying and selling after potentially climactic moves, which can be identified on the chart as large range with trend bars that often extend far beyond the edge. I believe that identifying true with-trend strength (or weakness, for shorts) and being able to discern it from exhaustion is perhaps the key technical skill of with-trend trading. No one talks about it very much, and it can’t be done perfectly, but, with some work and hard study, you can learn to dodge the most obvious bullets. (Note that this will also take you deeply into sentiment analysis and understanding crowd behavior.)

The actual entry trigger deserves a little bit of attention too. I’ve found a useful refinement is to use a lower timeframe breakout as an entry, and this can be as simple as buying a breakout of the previous bar’s high. You can also trade around previous lower timeframe pivots (for instance, a failure test), and here is where some of the most powerful multiple timeframe confluences come into play. Other traders will scale into pullbacks, but this requires a level of discipline that can be challenging for newer traders. Because you’re entering a market that, by definition, is moving against you, you will be “biggest when wrongest”, and this can be a problem if stops are not respected. Now, about those stops…

Stop location

Mark Fisher used to constantly remind traders he worked with and trained that the most important thing is knowing where you’re getting out if you’re wrong. Know this, on every trade, and respect that point, and you’re already far ahead of the game. The problem with pullbacks is that you are often entering a market that is not yet moving in the direction you want, so some degree of “slop” is required. We can be precise, but there are limits. Without belaboring the point, you will learn where to set stops, and they should go somewhere beyond the previous extreme. (I’ve covered this in considerable detail in one of the weeks of my free trading course.) As a starting point, 2-4 ATRs beyond the entry is a good, very rough guideline.

Trade management

If there is magic here, it happens in the trade management. I’ve found it helpful to take first profits when my profit is equal to my initial risk on the trade, and then to scale out of the remainder. This doesn’t work for those traders who hope to hit homeruns, but it certainly does drive toward consistency. There are other ways, but the key is to define your trading style and manage accordingly. I’m a swing trader, usually playing for one clean swing in the market. As such, I need to be proactive about taking risk off the table, but different styles will require different techniques.

Last, it is sometimes possible to use pullbacks in trending markets to build substantial positions. One idea that might be useful is to get in a pullback, take partial profits and hold the rest for a swing, and then get into a second trade if another pullback sets up in the same direction. Manage that new trade in the same way; the key is that you are always taking partial profits and moving stops so that you never have on more than a single trade’s risk. This is a smart and relatively safe way to “pyramid” into a trend move. Do remember that sometimes “stuff happens” and you may see a very large gap in a market, your stops may get slipped, and you may have a much larger than expected loss. Always respect risk, first and foremost.

Some examples

I thought I’d leave you with some examples. This idea of trading pullbacks is not just theoretical; it’s one of the most useful techniques I have found in years of trading. The following trades are trades that were identified in real time (many days and weeks in advance of entries) and published in my daily research I write for Waverly Advisors. (Obligatory plug: we’d love to have you take a look at our work!) Most of these recent pullbacks have been winners, and some gave way to outstanding trend moves, but I also included a losing trade here as well. Only entries are marked on these charts, but consider these in context of the ideas and guidelines I’ve given here. This is, truly, one of the simplest and most powerful trading patterns, and is useful to traders working on timeframes from intraday to months/quarters. If you don’t trade pullbacks, maybe you should?



Adam Grimes has over two decades of experience in the industry as a trader, analyst and system developer. The author of a best-selling trading book, he has traded for his own account, for a top prop firm, and spent several years at the New York Mercantile Exchange. He focuses on the intersection of quantitative analysis and discretionary trading, and has a talent for teaching and helping traders find their own way in the market.

This Post Has 12 Comments

  1. A_Joe

    Thank You ! Another Superb Post.

  2. irdoj75

    Thanks. Helps a lot, especially the comments about visual symmetry and “ugliness”.

    I only wonder why you choose/recommend such a wide stop. 2 ATR would be roughly the upper K-channel in your charts, 4 ATR double. When looking at e.g. the losing trade in ES I would not like to wait until hitting 1920 or even 1910.

    Also, the profit target moves away with a wider stop. Maybe I am wrong and oversimplifying, but looking at the examples I’d estimate 1 clean swing to be 2-4 ATR. Assuming that you will not get perfect H/L of the swings, you may be missing your profit target more often than not.

    Hey, I know you know how to do it – I guess I just reflected my own doubts and fears… :-/

    1. Adam Grimes

      It’s just what I have found works well for me… and that 1R stop is rarely hit. Of course, the tradeoff is that 1R profits are also difficult to achieve. I think you can work with tighter stops if you choose to.

  3. Jack

    I think that watching Adam’s past webinar would be helpful to understand his meaning of stop-loss placement. I remember his 2-4ATR stop is initial stop, which protect your from the movement of market noise. Of course, it will be better if this 2-4ATR stop is around the previous extreme level.
    But it seems contradict to the profit target. After some thinking, my understanding is that it relate to some philosophical thoughts about market behavior. The 2-4ATR stop is emergency rescue stop, if the noise move stops you out, the noise move also can give you a windfall profit about 2-4ATR. In the long run, this part movement will offset each other and level no real impact to your P/L.
    secondly, you do not have to wait you initial stop be hit. You can manage to get out based on the price action before stopping out. This part will match your part of managing to get out before your profit target based on the price action.
    Finally, if price move to the desired direction soon (your entry is confirmed), you then move your stop to new level ( actual risk).
    All of those above is based on the assumptions that you can read price actions well and your setups have real edges.
    That’s my understanding of Adam’s words. I am not sure if I am right. The finally explanation belong to the author of this post. ๐Ÿ™‚

    1. Markus

      Very interesting discussion.
      In the past my SL were only about 1-1.5 ATR away, since I tried to find an entry with minimum heat. But I have changed to a wider SL, because I have changed my entry process. In the past I have faded the pullback move and tried to enter around a 50% retracement. Unfortunately the best moves have sometimes only retracements of 30% or less and so I missed some pullback setups because I have not been aggressive enough. And on the other side some retracements were much deeper – 75% or more – and I got stopped out before the real trend leg set up.
      Adam has inspired me through his book and blog to use another technique. Now I enter after the pullback is over and I see a lower timeframe push in the direction of the new trend leg (if I cannot find a lower timeframe overextension entry). This has the advantage that I miss less setups (with a favorable context) and I have also confirmation but it has the disadvantage that I usually have more heat or at least have to plan for more heat and therefore my SL has to be farther apart.
      Another reason for a wide SL of 2-4 ATR is that you might want to plan for a complex pullback.
      IMO the most important part for the trade is the right context. If you have a favorable context for a pullback it is very important that you have an entry techinque which makes sure that you will not miss the trade and also that your SL is way out of the noise.

  4. steve goodwin

    I think the key part of proper stop placement, “where one is getting out when they are wrong”, that I think is difficult to master has to do with one’s correct determination of the stop placement in relation to the timeframe of the trade they are entering.

    Identifying the proper location of “wrong” on an intraday timeframe trade using a 15 minute chart setup is much different than the “wrong” location for a swing trade using a daily chart setup.

    1. Adam Grimes

      It is certainly one key, but I think consistency is important. Similar ratios (i.e., visual distances on charts) can be used on different timeframes, but, yes, there are some adaptations needed for intraday data.

  5. gb

    When stocks gap or have large morning moves then Keltner channels (or similar) take too long to catch up to be useful for intra-day trading.
    What other approaches could be used to identify a pull back somewhat objectively?

    1. Adam Grimes

      There are a number of ways to deal with this, but I usually just ignored the indicators until they settled down after the morning gap. Could also use premarket data or code an indicator that adapts to the gap.

      It’s not so much that the indicators aren’t useful, it’s just that you have to figure out how to adapt for that first hour or so on big gap days. Many possible solutions–just find one that works and be consistent.

Comments are closed.