Let’s face it. By now, most traders have learned (often the hard way) that the best way to make money in the market is to do the opposite of what you want to do. Unfortunately it is human nature to want to follow and trade in the same direction as the market is moving. For example, we’ve all seen a big up day in the market where we got excited and went on a buying spree, only to see the market reverse the next day and put our positions in the red. Similarly, we’ve all seen a big up day in the market where we were not excited and did not buy anything, only to see the market continue moving to the upside over the next several days.
No doubt, many traders have been burned too often by following the market and subsequently give up. The market can truly seem to be out to get you sometimes.
For short-term swing traders, knowing when to buy strength and when to sell it will determine whether they succeed or fail. Sure, good position-sizing may mean they don’t lose all their capital, but unless these traders figure out how the market really works, they will eventually give up.
I’m here to give you the secret ingredient to the market recipe–the one that will help you know whether you should be trusting or fading a market move. This secret ingredient will help you know when to buy strength, and when to sell it. In short, this ingredient should be as important as your best indicator, your favorite technical pattern, or your favorite strategy. In essence, this ingredient will tell you when to do the opposite of what you really want to do.
Before I give away the secret, as with any ingredient, in order to use it wisely, you need to understand how it works and when to use it.
A Short Primer- Daily Follow Through and Mean Reversion
I’m highlighting Daily Follow Through (DFT) because it is a very basic yet robust method of figuring out whether a market is following-through or mean-reverting. While there are indicators which can describe/measure follow-through and mean-reversion, I believe this is the simplest method and therefore the easiest to understand and reproduce.
DFT describes the ability of the market to follow-through on a move (some may call it “trending”). If the market closes higher on Monday and then again closes higher on Tuesday, it has followed-through on Monday’s move. This is the opposite of Mean-Reversion (MR) which occurs when the market revisits prior price levels before moving again with the primary trend (This is admittedly an over-simplified definition of MR). A mean-reverting market will have the tendency to close lower on Tuesday after a higher close on Monday.
Over the past 15 years or so, any trader relying solely on DFT as a strategy has lost all of his capital. Since the mid 1990s, MR has been the driving force behind the U.S. markets. However, there are times when MR doesn’t work very well, and traders will need to adapt to a market environment which is based on DFT.
Part Two- Looking at the Results of a Simple DFT and MR Strategy
In part two, I’ll provide some statistics as well as chart porn for a simple DFT and MR strategy. We have to first examine when each worked and then compare the results to the broader market environment. Once that is complete, we’ll be very close to understanding the secret ingredient.
tease.
and yeah, i will be back for the secret ingredient.
good stuff from you, as always.
…and i especially like this gravatar i did not load. very nice default!
I agree. It is one of the best I have seen yet!
Thanks! Sorry about the tease…SNL was getting ready to come on and I wanted to watch it!
looking forward to the rest of this.
Great post ..looking forward to the rest of it.
If you can keep the clarity and simplicity of the rest, at this same level, Fly should grant you the award of professor of the year.
Hawaii, no problem on the simplicity. I prefer to keep it simple. Often, in the market, simple outperforms complex.
Hi, I have tested many trading systems using tradestation/ninjatrader or others. What I found is the most robust systems are all trend following trading system. It is EXTREMELY difficult to design a robust mean-reversion trading system. If you don’t optimize mean-reversion system, its performance is very poor. If you want good performance, you need to (over)optimize it. But it is very susceptible to over-optimization/curve-fitting. The parameters are very sensitive, if you change the parameter somewhat, its performance may reduced dramatically. So I begin to doubt the robustness of mean-reversion systems. Obvious its robustness is not in the same league as trend following. You can easily design a trend following system with the same parameters and apply to all future, currency markets for over 20 years and still make good profit. That is extremely difficult for mean-reversion system, I guess after each 5 years, the market rebound/mean-reversion patterns changed, so we need to reoptimize mean-reversion system every 5 years or even more frequently. Another big disadvantage for mean-reversion system is it may have a big drawdown each trade and it can’t be avoided. Experiment has shown mean-reversion system needs to use a much wider stop loss than trend following system. A trend following system can use 1 to 1.5 times ATR as stop loss without affecting the performance much. But use a tight ATR as stop loss for mean-reversion system will kill its performance. From my testing, the stop loss for mean-reversion systems need to be at least 3 times ATR or even higher without affecting its performance much. So for each trade, its risk/stop loss is much higher than trend following system though trend following has a much lower winning percentage. Maybe the reason is I didn’t find a robust mean-reversion system, but I have tried RSI, Bollinger band, z-score and add ADX, moving average as filter, or only trade with the direction of big trend, I have tried everything I can think up and I see in books and magazine. The best I find now is to use z-score and only trade with the direction of big trend, and add the slope of moving average as filter, that is the best I can find now. But the performance is very sensitive to the length of that moving average. So I doubt the robustness of all mean-reversion systems or maybe just because of my ignorance of mean-reversion systems.
Jing, thanks for the comment. Respectfully, much of what you have written is not true in my testing.
I would guess that you are not using de-listed data when testing your trend following systems?
Really, you are talking about the inherent differences between short-term and long-term systems. They are different animals entirely.
A tight ATR stop will not kill a MR system. Yes, it will, if it doesn’t trade very often. However, if it generates lots of trades, a tight stop will work very very well. Short-term MR systems have high win rates, typically between 60-70%. Thus, even with tight stops, they perform well.
Are you testing position-sizing as well?
The power dip uses a variety of stops, but 3ATR stops are currently averaging out to be stops between 5-8%. I would bet you would consider those tight stops. Yet the Power Dip model using 3ATR stops is doing well, even though many of the stops it has been using are less than 10%.
I would like to hear your thoughts. My guesses are that you are not using de-listed data and are not using percent-risk position sizing in your tests, and that your MR systems don’t trade very often.