Sunday, November 12, 2006

Why It's So Difficult To Be A Trend Follower

The excellent Barchart site contains a great deal of performance information about stocks, sectors, ETFs, etc. Here is my little mashup of some of their findings:

Take a 40-day version of the Commodity Channel Index. When the oscillator moves above 100 (showing good upward price momentum), buy. When it moves back below 100, sell. Similarly, when the Index moves below -100 (showing strong selling momentum), go short. When it moves back above -100, cover shorts.

Let's see how this little system performed for the S&P 500 Index (SPY) over the past two years:

* The system made 48 trades, averaging 7 days per trade. We had 11 winning trades and 37 losers.

* Overall, the system lost a little more than 18 SPY points (the equivalent of 180 points in the futures). This, of course, occurred during a bullish period of market prices.

* The average size of the winning trades was considerably higher than the average size of the losers, but the better than 3:1 ratio of losers:winners kept the system in the red.

Ironically, this might be a great nucleus for a trading system. It's telling us that the market is trading in a countertrend mode. If you had a method to limit your losses and faded the system signals, you'd have a fighting chance to make some good money. After all, a consistently losing system *does* have a potential edge.

Now consider the following:

If we traded that exact same system in GOOG, we'd have had 33 trades averaging 11 days per trade. We'd have had 14 profitable trades and 19 losers, but the system would have made over 134 GOOG points over the last two years by catching trending moves. Almost half of that profit came from a single big trending move.

The examples show how hard it is to be a trend follower. The indices are reversing directional moves, making trend following a losing proposition. Even with good trending stocks, you have to be able to tolerate frequent losing trades. Most of all, you have to be consistent enough to stick with the system for the occasional huge winner. There *are* traders able to succeed with trend following, but shrewd stock/asset class selection and steely discipline are necessary.

4 comments:

voodster said...

All due respect, but the simple rules you used to simulate a trend following system is not exactly a trend following system.

Brett Steenbarger, Ph.D. said...

Hi Voodster,

You're right; a moving average system might better serve the purpose. The example I used might better be called a momentum trading system. The dynamics of wins and losses, however, are quite similar between the two, with many small losses and fewer, but larger gains. Thanks for the observation--

Brett

cpptrader said...

Hello Dr. Steenbarger,
I have been toying with a similar type system - and you are absolutely right that asset [class] selection is essential. You mentioned that "a method to limit your losses" is required. In a counter-trend trade, what sources do you know of to help me develop a good limit system. The problem is that when I place the trade against the move, I have had a win rate between 70 and 80% of the time, but the position often initially moves against me, with those big moves, like the one you mention in GOOG, being the ones that hurt. I have not formalized ideas into tests yet, but have thought of time and loss based limits, etc. - trailing stops do not seem to be effective. Any thoughts?

Thanks for the postings,
Matthew

Brett Steenbarger, Ph.D. said...

Hi Matthew,

I find that, in a counter-trend trade, I am best off to wait for buying or selling to dry up first (fewer large traders lifting offers or hitting bids) and then try to ride the move. I have a few blog posts on the topic and use the Market Delta program to track volume at the bid vs. offer. I also find that waiting until the distribution of the NYSE TICK moves my way helps with entries. I'll try to illustrate these methods in my next morning trading session.

Brett