One of the decisions that every trader must make is selecting a timeframe for their system. While this may seem like a simple decision, there are actually a lot of working parts involved. Systems that trade on longer-term timeframes are more likely to catch big chunks of trends while posting lower commission costs. On the other hand, short-term time frames will allow more opportunities for a system that is profitable on a per-trade basis.
Deciding on a timeframe also brings into question an individual trader’s commitment. Trading a short-term system can require quite a bit of regular time investment from the trader. Longer-term systems can be more useful to traders that aren’t able to commit to spending a large portion of every trading day in front of a computer.
The best way to make this type of decision about your own system is to backtest it on different trading timeframes to see how it would have performed. You can then compare these results to how trading that system would have fit into your lifestyle over that time. Would the returns have been worth the investment?
Simple Moving Average Systems
One of the simplest systems you could possibly trade is simply buying a general market ETF when the price closes over a certain simple moving average (SMA) line and then selling it when the price closes below that line. In order to create some comparison data, I backtested this system on the Vanguard Total Stock Market EFT (VTI) from January 1, 2001 through December 31, 2010.
For each of these backtests, I used a starting capital of $10,000. In addition to exiting on a close beneath the SMA line, I also used a 10% trailing stop on each trade and assumed a $7 commission on each trade. In order to compare different timeframes, I ran three separate tests using the 50-day, 100-day, and 200-day simple moving averages.
50-day SMA System
The 50-day SMA System ended the ten year testing period with $13,430.34, which was good for a total return of 34.3%. The system logged 17 winning trades and 58 losing trades, which works out to a winning percentage of 22.7%. It had a maximum drawdown of 35% and paid a total of $1,057 in commissions.
100-day SMA System
The 100-day SMA System ended the testing period with $15,878.26. This represented a return of 58.8% over a ten year period. The system posted 9 winning trades and 35 losing trades, which means it had a winning percentage of 20.5%. It had a maximum drawdown of 23.9% and paid a total of $623 in commissions.
200-day SMA System
The 200-day SMA System ended the test period with $17,246.01, which was good for a 72.5% return over the ten year test period. During that time, the system only recorded 5 trades. Of those trades, 2 were winners and 3 were losers. This represents a winning percentage of 40%, but that could be the product of a small sample size. This system posted a maximum drawdown of 34.2% and paid a total of $77 in commissions.
Simple Crossover Systems
One of the first systems I wrote about was the 10/100 Moving Average Crossover System. I tested this system on the same ETF using the same conditions as the previous tests. Then, I tested a shorter-term system that traded 5/50 crossovers and a longer-term system that traded 50/200 crossovers.
The 10/100 Crossover System
The 10/100 Crossover System ended the ten year testing period with a total value of 18,978.69. This was good for a total return of 89.8%. During those ten years, it logged 7 winning trades and 9 losing trades. This represents a winning percentage of 43.8%. The system recorded a maximum drawdown of only 16.2% and paid a total of $231 in commissions.
The 5/50 Crossover System
The 5/50 Crossover System ended the testing period with a total value of $16,075, which was a total return of 60.7%. The system posted 13 winning trades and 22 losing trades, which was good for a winning percentage of 37.1%. It suffered a maximum drawdown of 22.8% and paid a total of $497 in commissions.
The 50/200 Crossover System
The 50/200 Crossover System ended the testing period with a total value of $18,095.29, which equals a total return of 81%. The system logged 4 winning trades and one losing trade. This works out to a winning percentage of 80% in what is obviously a small sample size. The system logged a maximum drawdown of 14.5% and paid a total of $77 in commissions.
Analyzing Our Results
As you can see, in both of these cases, the longer-term systems are able to match or outperform the shorter term systems. It is very interesting to note how the longer-term systems are able to do this while making far fewer trades, which results in much lower commissions costs.
It is important to keep in mind that both of these systems are trend following in nature. That explains why they perform better on a long-term basis. We would likely get very different results if we ran this type of experiment on mean reversion systems.
This simple distinction is a huge lesson that we can learn from these tests. If you want to trade on a long-term timeframe, you might want to consider a trend following system. If you are trying to trade on a shorter-term timeframe, a mean reversion system might be a better fit. You should do a careful self-analysis to decide which style and timeframe is best suited for your personality.
Scott says
Hi Andrew,
There are some folks who believe that each time frame will “trend” with similar frequency —- relative to its own scale.
Andrew Selby says
Hey Scott,
That’s a great point. Perhaps a future post idea!
MTLibrary says
Larger time frames can be good for overall bias, but the core framework of any strategy should be built on one of only a few select frames. I’m speaking about those that are consistent and the same from broker to broker. For example: GMT+0 (broker 1) and GMT+1 (broker 2) will not have the same bar OHLC on a H4 candle. The same can be said for a daily candle, and so on. It’s important to see what the rest of the trading world is seeing – therefore one should only use time frames that are the same everywhere regardless of your broker’s time zone – which they can change, thus giving your strategy entirely different results. Consistency is key!
Andrew Selby says
That’s a really interesting point! Thank You!