Exits & Stops
The strategy that we’ve developed so far in this series is technically a day-trading strategy, because it exits all positions on the same trading day that they were entered. For the sake of illustration, however, let’s assume that you’re willing to hold onto your positions overnight.
In the previous section we mentioned that the strictness of your exit rules can affect the profitability of your strategy. Whereas stricter entry rules tend to limit the number of trade signals that are generated, stricter exit rules tend to increase the duration of the trades. As a simple example, let’s consider two simple exit methods that could be used for long trades:
- Close > 3-day moving average of closing prices, or MA(3)
- Close > 5-day moving average of closing prices, or MA(5)
As a general rule, when a stock has pulled back its price will cross above the MA(3) before it crosses above the MA(5). Therefore, we consider Close > MA(5) to be a stricter rule than C > MA(3) because it is more difficult to achieve (occurs less frequently). Another way to state this is that if we compare two strategy variations with identical entry rules, the variation that uses the MA(3) exit will have shorter trade durations than the one that uses the MA(5) exit. Let’s see how this might affect the results of our sample strategy.
For simplicity, we’ll just look at results for the long side of our strategy. The table below has one new column, Avg Days Held. This measures our average trade duration in days. We have placed the MA(3) variations next to the MA(5) variations for easy comparison.
For any variation of entry criteria, you will see that the average trade duration is approximately a day and a half longer when using an MA(5) exit than when using an MA(3) exit. The Average Gain per Trade is also slightly higher with the stricter exit rule. Of course, there are limits to this approach: if you make your exits too strict, the stock price may roll over and start to erode your gains. Experimenting with a variety of exits will allow you to judge which criteria fit best with your overall trading plan and style.
A topic closely related to exit methods is the concept of stop losses. There are many ways to set stop loss orders. A few methods that pertain to long trades are:
- Exit if the price falls X% below the entry price
- Exit if the price pulls back $Y (or Z%) from the highest price (trailing stop)
- Exit if the price falls below the N-day moving average
- Exit if your loss on the trade exceeds some threshold amount
For some traders, stop losses bring peace of mind, because they limit the amount of money that can be lost on a trade, although it’s important to note that a stop loss order does not guarantee that your loss will be limited to the amount you intended. In general, however, stop losses work pretty much the way you intended them too.
The problem with stop losses is that they generally decrease returns. Over the years we have run many tests that include stop loss orders, and those tests consistently show that stop losses decrease the average gain per trade. This is particularly true of mean reversion strategies, where a drop in price (i.e. a bigger pullback) actually represents a bigger opportunity, not a failure of the central thesis of your strategy. Let’s look at an example.
Below is another side-by-side comparison. On the left side are the results that we examined previously that utilize an MA(5) exit. The results on the right side are the exact same test, but with a 2% stop loss added. In other words, if the price falls 2% below our entry price, we exit the trade immediately. Notice that for every variation of our strategy, using a stop loss results in a lower average gain per trade. In fact, we lose anywhere from 64% to 83% of our profits by using this stop rule! Additionally, our win rate gets slashed by at least a half, and sometimes closer to two thirds.
The Effect of Stops
The general rule of thumb is that if placing stop loss orders allows you to sleep better at night, then go ahead and use them. Just make sure to also incorporate them into your back-testing so that you have a good grasp of how much this type of “insurance” will cost you over the long run.
You now have all the tools necessary to develop your own quantified trading strategies. Over time you will learn which techniques tend to produce good results, and which ones have little effect. You will also determine the types of strategies that are best-suited to your personality and trading style. And finally, knowing how strategies are developed will likely give you a deeper understanding and appreciation for high-quality strategies created by others, and also to more quickly spots flaws in the logic or methodologies of traders who present strategies of lesser quality.