Where to Put a Stop Loss
It is virtually a "law" of trading in the stock market that wherever you place your stop loss, it will occasionally be triggered by a stock just before it resumes its climb to higher levels. That is just something to be expected if you use any stop-loss. Unfortunately, not using a stop-loss is asking for trouble of a much greater magnitude, and the market loves to reward the foolish, lazy, or stupid, with the just recompense of their behavior. It makes no difference if you set the stop at 10% or at 3% from the low, high, or close. You can use stops that are volatility-based, use Fibonacci retracement ratios, Gann analysis, pivot points, percentage declines, or any other approach. No matter how sophisticated your mathematics is, you will often find you have sold for no good reason other than the occurrence of a temporary price spike that was just sufficient to trigger a stop loss -- your stop loss. Learn to live with it..
On the other hand, you can control risk and have some say about the probable frequency with which you will be ejected from a position because of such a spike. The further your stop is from recent price action, the less likely it is that it will be triggered. However, the further your stop is from the price action, the more risk (downside price excursion) you are going to have to tolerate. Not using a stop at all means you are willing to accept unlimited risk. Using a "tight" stop means you are willing to tolerate very little risk but you dramatically increase the chances that even a minor spike will eject you from the position. The tighter your stop, the more ejection-causing spikes will occur in any given time period. The only way to resolve this dilemma is to find the best tradeoff between an acceptable frequency of unnecessary ejections and an acceptable amount of loss that you incur because of that ejection. In other words, you must find the compromise that induces the least amount of pain (psychological or financial).
Magee and Edwards (Technical Analysis of Stock Trends) teach that a good stop based on closing prices is one that is placed 3% below a rising trendline. The stop is triggered only if the stock closes at or below the stop. However, if a trader intends to sell on the basis of intra-day price activity rather than on the basis of closing prices, they suggest that the stop be placed 6% below the rising trendline. Below the trendline or below the most recent minor dip is usually the best place for a stop. However, sometimes there is no trendline or obvious recent minor dip. That is when you must use a mathematical stop. Either a simple percentage based on the highest high, low, or close since you purchased, or a volatility-adjusted variable stop placed relative to the highest high, low, or close since you purchased will serve the purpose. Magee and Edwards, Weinstein, Schwager, Murphy, and many others use trendlines, dips, or moving averages as a reference for placing a stop. Rising trendlines follow the lows, dips are nothing more than significant recent lows, and moving averages generally follow a rising stock somewhat below its recent lows. Therefore, it also makes sense, in the absence of all of these, to use the recent highest low as a reference for placing stops. With no trendline or dip to use as a reference, you could simply place your trailing stop 3% or 6% (or some other distance) below a moving average that closely follows the trends of significance to you, or even below the highest low achieved by the stock since your purchase.
The anticipated average holding period has a very big impact on how tight your stops are going to be. For example, the "sweet spot" for 10 to 15 days might be about 2.5% (it really depends on the volatility of the stock). The short end of the "swing-trader" spectrum is about 3 days or less (a large number of traders focus on holding periods of up to about one week) and the long end of the spectrum is 8 to perhaps 10 weeks. The remaining swing traders focus on the time frames in between. At the very short end, 2.5% would allow too much of a decline relative to the expected gain. However, it works well when you are trying to lock in a two-week move involving a 4% to 10% gain. [When I say "it works well," I mean for a straight percentage stop loss. Of course volatility-adjusted stop losses are far superior. Straight percentages may be easy to calculate, but volatility-adjusted stop losses are less likely to be triggered by a meaningless price spike. All comments here about straight percentage stop losses should be read with this context in mind.] If the stock is not too "wild," it will also work beautifully for moves of a month or more to lock in gains of 10% to 20% or even more. However, you may have to loosen the stop a little for more volatile stocks and for regular holding periods of more than 15 days. For longer-term investing, for example, a stop that is up to 6% below the highest low reached by the stock since it was purchased can be very effective. A stop that one stockdisciplines.com trader experimented with and found to be very useful for intermediate-term trading is one that is set 4% below the highest low. In use, it was infrequently triggered by a whipsaw and it did not give up much of the gain of the bigger moves (relative to other straight percentage stop losses). However, it would also give up 4% or more of the smaller 8% moves. That is why some traders focus on stops of 3% or less below the highest low (below the highst high to capture the most gain from a surge). The tradeoff was the greater frequency with which a person is needlessly stopped out of a rising stock. It would be best if you worked out a personal stop-loss system, one with which you can be comfortable.
Walk away from stocks that often spike down. The specialist may simply like to "gun" the stock in order to take out the stop-loss orders waiting at the lower prices. That is, the specialist temporarily drops the stock price to trigger the sell orders associated with the stops so he can buy those shares at the lower price and sell almost immediately afterwards at a slightly higher price. When considering the purchase of a stock that often spikes down, the trader should try to place the stop just outside the specialist’s spiking comfort zone. If such a placement requires the assumption of too much risk, find another stock. I prefer to concentrate on stocks that rarely spike. Look at charts and notice the length and frequency of downward spikes. Try to determine the percentage drop these spikes represent.
A volatility-adjusted stop has a more universal application than a simple percentage stop because in addition to adjusting for volatility it also adapts to the time period of the particular "analysis unit" used (15-minute price bars, 30-minute price bars, daily price bars, and so on). Rigid percentages cannot do either, but they are easier for the non-mathematician to calculate. Straight percentages are a "ham-fisted" approach to stop losses. They are relatively crude and are more apt to be triggered unnecessarily than volatility-based stop losses. If you are trying to compute your own stop-losses and you do not have a mathematical background, you might do well to use a stop loss calculating tool We make available such a tool because we believe that calculating a volatility-adjusted stop loss is a more elegant, accurate, and far more logical approach to finding a stop loss that has a low probability of being stopped out of a position because of a meaningless price move.
The stop is not necessarily your sell discipline. However, it is definitely your safety net. It will preserve assets if you are not paying attention to your stock’s behavior during the day. If you do not have time to be riveted to the etchings the stock market makes across your computer’s screen, you only need to take a few minutes to go through your positions once a day (even while the market is closed) or once a week (even on weekends) to set your stops. Then you can ignore the market until you make your next stop adjustments. However, if you happen to be watching your stock and it does not "behave" like it should, simply remove the stop and sell the stock.
Dr. Winton Felt maintains a variety of free tutorials, stock alerts, and scanner results at www.stockdisciplines.com has a market review page at www.stockdisciplines.com/market-review has information and illustrations pertaining to pre-surge "setups" at www.stockdisciplines.com/stock-alerts and information and videos about volatility-adjusted stop losses at www.stockdisciplines.com/stop-losses