Some markets exhibit trending behavior while others do not. I was wondering what would be a good way to determine if a given market exhibits trending behavior. One simple method to accomplish this is to build a simple trend following strategy and test how well it performs. Then build a simple mean reversion strategy and apply it to the same market. From there we can see which trading strategy performs better. This simple trend following strategy consists of a single 50-period simple moving average (SMA) on a daily chart. To keep things simple, the system only takes long signals. It opens a new trade when price crosses above the moving average and closes that position when a daily bar closes below the SMA. I’m not attempting to create a trading system per se, but creating an indicator that measures a market’s trending characteristics.
- Daily Bars – No Commissions – No Slippage
- Buy close of bar when Close > SMA(50)
- Sell close of bar when Close < SMA(50)
Below is the equity graph created with this system on the S&P E-mini futures market from September 1997 to July 2015. As you can see the equity curve remains in negative territory most of the time and is unable to make any substantial gains.
We can now run the same strategy on the Euro currency futures. Below is the equity graph on the Euro from May 2001 to July 2015. Notice anything different? The equity curve is climbing higher and higher. Only over the last couple of years have things changed.
By creating a simple trend following system that utilizes a 50-period moving average, I hope to demonstrate that the S&P E-mini (ES) futures market can be unfavorable to trend following systems. At least, for short-term trending systems. On the other hand, the Euro currency futures shows much stronger trending characteristics. At least until very recent times.
But what about longer lookback periods beyond 50? We all know the S&P can be in a persistent bull trend for years as we have seen for most of the years since 2009. I’ll use TradeStation’s optimization feature to test lookback periods from 10 to 200 in increments of 10. This will generate a bar graph of profit vs the lookback period. Here it is:
We can clearly see that lookback periods of 50 or less produce losing results. On the other hand, values above 50 generate profitable bars. In general this seems to indicate that the S&P market has a mean reverting tendency on shorter periods. In other words, longer-term trend following on the S&P may work better than aiming for shorter-term trending following systems.
For a further demonstration of trading the S&P on a short-term basis, see this article written years ago called, Better To Buy Strength of Weakness? In this article I demonstrate that short-term strength on the S&P futures market is often followed by weakness over the next five days. Likewise, short-term weakness tends to be followed by strength over the next five days. So, this study and the one presented in this article tend to reinforce each other. It’s these types of observations that can be used as a starting point for building a system.
In very general terms, if your trading the S&P futures market with holding periods well under a few days, you may want to focus on mean reverting strategies. If you’re looking to build trend following systems, you probably should be looking at longer holding periods well beyond five days.
What might be interesting to check out is if momentum reappears as dominant in the intraday time frame.
In theory, every trade is a momentum trade of some sort, since once you’re in a trade, you hope it continues in the direction you planned. The difference is that a mean reversion trade gets in before (ideally) the price changes direction. Momentum strategies get in once the trend is established. But they both want the trend to continue rather than wander.
So..at what time scale does momentum come back into play for the E-mini? Just curious.
Matt, that’s a good idea. I’ll think about it and add it to my to-do list.
If we look back Jeff, S&P500 seems to behave in a trend following regime before late nineties. It’s not easy to detect a change in regime, but when the trend following or mean reverting regime changes, this change lasts for years. Once the new regime has set up, traders have years to exploit it. Only this seems to be incouraging.
And when it does. It would be great to understand why such changes happen.
https://nightlypatterns.wordpress.com/
Agreed. If I remember correctly after the dot-com bust of 2000, the S&P changed.
Marco–
“…when the trend following or mean reverting regime changes, this change lasts for years.”
On what data do you base that claim?
Mark