Thursday, June 21, 2012

The Art of the Stop/Loss

Crack any book on trading, and you'll soon come across a statement that goes something like this: Set a stop/loss on your positions in order to limit your losses.

I've read that statement and its cousins so often that it has achieved the status of indisputable truth, like the falling of apples when dropped from leaning towers and the ethical strictures against coveting neighborly goods and spouses.

However, one lesson I've learned in thirty years of trading is that the indisputable must always be disputed. Could it be the authors of all those books on trading are wrong about stop/losses?

I've come to the subject as a result of a bad trade. I set a stop/loss too close to the entry price, it inevitably got taken out, and the trade was closed, for a small loss. I wrote about in my post-mortem here.

In that simple tale of a minor failed trade lies a host of assumptions: Stop/losses are necessary. There must be rules for setting them. When a stop/loss is taken out, it represents a failure because it creates a loss, a cost to the trader.

Stop/losses are necessary.

But are they? A stop/loss, after all, represents doubt in the trader's analysis. A stop/loss says, "Yes, I entered this trade, but I'm not real sure about it, so I'll fudge."

That strikes me as being a pusillanimous stance. If I'm uncertain of my analysis, why in the world would I open the position? Ought I not be confident that I've done my job right before committing my funds?

By that measure, a stop/loss is a symptom of weak analysis. It also points toward the evil of habit of trading just because that's what a trader does -- trading as an addiction, an illness that preys on all of us who love the drama, the uncertainty and the head rush of the markets. 

A stop/loss is also a means to avoid paying attention to a trade. The good trader will know at all times what the positions are doing. If the analysis was wrong wrong, then the trader will use judgment about whether the time has come to close the position.

A stop/loss can be represented as a judgment, true. But it's a judgment made days or weeks or months earlier. Trading is something that happens now, and it is in the Now that the best trading judgments are made. If I'm going to close a position, I need to do it based on current information, not on old information.

In the markets, sometimes five minutes ago can be old.

Argue though I might, it is a fact that stop/losses are a universal practice. And there are practical reasons for doing so. The main ones are speed and time.

Sometimes the markets can crash very quickly. The Flash Crash of 2010 comes to mind. And I have vivid memories of Black Monday in 1987. A stop/loss in either of those circumstances could have gotten me out of falling positions well before I could come to a decision and intervene.

In 2010, that would have played into a head fake, and I would probably have re-opened my positions the next day. In the case of 1987, a stop/loss would have spared me months of painful loss.

Less dramatically, a lot of traders work, and the boss takes an unfriendly attitude toward using company computers to trade on company time. Many in the worker-bee community place trades in the evening for execution at the market open. A stop/loss can be important toward safe-guarding a position when the trader can't trade for extraneous reasons (like mean bosses).

There are rules for setting stop/losses

A standard recommendation from trading coaches is to set the stop/loss at 3% below the entry price.

This is inevitably, and quickly, followed by the admonition that the trader must determine the level of potential loss that they can handle, a caveat that I find to be most unhelpful. I mean, for a lot of people, any loss is hard to handle. Others are willing to hold on to a stock and see it drop to zero before they're willing to part with it.

Essentially, the coach is saying, "Three percent, but don't hold me to it. Come up with your own rule."

If setting a stop/loss is a serious recommendation, then there ought to be a rule on setting it. And there isn't. Another recommendation I've seen is to set a 5% stop/loss. I've even seen 1% offered up on occasion.

Fixed percentages, of course, ignore differences behaviors among stocks.

Some stocks move in wide range each day, such as Netflix (NFLX), a $66.50 stock that has a daily average movement of $3.16, or 4.8%. Odds are good that a 3% stop/loss on NFLX will be hit, on average, every day.

Others move in narrower ranges, such as Johnson & Johnson (JNJ), a $66.27 stock that has a daily average movement of 78 cents, or 1.2%. A 3% stop/loss on JNJ would be far less likely to be taken out that that on NFLX.

An alternative to fixed percentages is to use the average daily movement to calculate the stop/loss. The Average True Range (ATR) gives a 10-day average movement. A stop/loss can be set as whatever multiple of the ATR the trader wants: Usually the ATR itself or double the ATR or for more leeway.

So in the case of NFLX, using the one ATR would produce a stop/loss of $63.64. For JNJ, that method would give a stop/loss of $65.49.

The most rational method, in my view, is to set the stop based on the trader's reasoning when entering the position. Generally, my reasoning goes something like this: The price is in an uptrend because it has broken above Level #1, and if it falls below Level #2, then it will no longer be in an uptrend.

Rationally, I would want to set my stop/loss at the point where the uptrend disappears, Level #2. This approach treats every position as a logical proposition. When events have proven the proposition to be false, then there is no longer justification for the trade and the position should be closed.

Another way of stating this method is to say that stop/losses should be set below support.

Perhaps using the price for the stop/loss is the wrong approach. There are many risks in life. When I'm driving a car, I have the risk of having an accident. I don't quickly stop my car and get out of it when I see the risks rise in heavy traffic. Instead, I carry collision insurance to mitigate the loss in case of an accident.

In trading, buying insurance on a position is called hedging. There are many ways to hedge, generally using stock options. The simplest is to buy a put option on a bull position, since it will gain money if the position loses. More complex positions, such as covered calls and option spreads of various sorts are nothing but hedges, intended to limit loss.

I've found as a trader that it's not difficult to judge the direction of a stock's movement. If I expect a stock to rise, it generally will. Eventually.

That "eventually" means that I'm horrible at guessing when that rise will occur. I count count that the times that I've opened a bull position, had it immediately start to fall the next day, closed the position for a loss, only to see it rise above my entry point.

That being the case, perhaps stop/losses are best set using the calendar. If I'm confident that stock will rise, by this argument, I ought to open my position and hold it for a fixed period of time -- a month or three months or a year, depending upon my time horizon. At the end of that period, I close the position, whatever the price.

A variant is to sell the position if it's unprofitable, but to hold it for a new period if it is making money. In any case, using this method I would sell in the middle of a period.

I've seen studies concluding that selling by the calendar, with a holding period of three months or so, is more profitable than relying on moving average crosses or other price-based strategies. (See page 143 of Way of the Turtle, a book about the trend-following Turtle Trading method. The book is well worth reading on many levels.)

Stop/losses are a failure when taken out

Any method used to protect a position from loss has costs. That's because any such method amounts to insurance, and insurance always carries a premium.

In the case of a simple stop/loss, the "premium" comes when the trade is stopped out below its entry level. The loss on the trade is the premium.

Of course, once the stock has moved away from the entry price in the direction of profit, then the trader will change the stop, making it less likely that it will produce a loss if taken out. That means free  insurance, always a good thing.

Hedges are never cost free. The cost of hedging a bull position limits on potential gains, as in covered calls and all spreads. Or it can be in the need to put up huge amounts of margin to ensure you can cover the hedge, as in a synthetic call or synthetic future. Or in the case of a put or call bought for insurance, it's the cost of buying those options.

The fixed period trade has "premium" similar to a price stop/loss. If the position is loss-making when the period ends, then that loss is the cost of insurance.

But is the cost of insurance a failure? Surely the premium paid for collision insurance on a car or life insurance represents prudence, not failure. It's a cost of living in an uncertain world. The same can be said for the various premiums used to mitigate loss in the markets.

No one could reasonably say that collision insurance failed if there's a crash, or life insurance failed when there's a death.

In those cases, there's a payoff, and in the case of a stop/loss that has been taken out, there is a payoff as well, in the form of losses not suffered because the stop/loss closed the position before greater damage could be done.

By that reasoning, when a stop/loss is taken out, it is an act of beauty and a tribute to the trader's forward-thinking prudence.

What I do in my account

I'm all over the map. For bullish options positions that gain value as they near expiration, such as covered calls and diagonal spreads, I tend to buy puts to insure the position and don't set price stop/losses. The options themselves have expiration dates that provide a fixed-term stop/loss for the position.

(Arguably, using puts to insure a hedged position with a fixed-term stop/loss -- an options expiration -- is overkill, and I'm considering discontinuing the practice on individual trades, instead buying deep-out-of-money puts on the S&P 500 index to insure against systemic risks.)

For option positions that are prone to time decay (the loss of value as the options near expiration) and also for stocks, I tend to set stop/losses based on the support levels -- my trade has a rationale based on the chart, and if the chart changes so as to invalidate my reasoning, then I close the position.

This can often be a very subjective process and lead to whipsaws. A chart has many levels of movement, from the slow and grand to the fast and miniscule. Selecting which level to trade is a subjective art.

As with everything in the markets, stop/losses are complex. There is no single "right" method in my opinion. What's important is that any method be selected after thinking through the various arguments, rather than simply accepting some else's word.

As a trader, I'm responsible for my trades. To accept someone else's "3% stop/loss" as gospel diminishes my responsibility.

Tim Bovee, Private Trader tracks the analysis and trades of a private trader for his own accounts. Nothing in this blog constitutes a recommendation to buy or sell stocks, options or any other financial instrument. The only purpose of this blog is to provide education and entertainment.
No trader is ever 100 percent successful in his or her trades. Trading in the stock and option markets is risky and uncertain. Each trader must make trading decision decisions for his or her own account, and take responsibility for the consequences.

No comments:

Post a Comment