A simple moving average system that still works!

OpenMind

Active member
Messages
130
Likes
12
I was playing with Excel over the weekend and thought I'd test a simple strategy on a selected market over a long period of time. The strategy I chose is a very simple one: one of the oldest and most basic trend following strategies:

Go long at the market if the price closes above 50-SMA + ATR(10)
Go short at the market if the price closes below 50-SMA - ATR(10)
Trail the stop on the 50-SMA.

I then coded it in Excel using EOD data of SMH (semi-conductors ETF) and results are very surprising. With a starting capital of $10,000 and 2% risk per trade, it has returned a cumulative total of 85% over 5.5 years, with the ending capital being $18,500 (unless I have made a silly mistake somewhere). In the intervening period, this market has seen bull, bear and sideways phases.

See the spreadsheet attached. You can enter the following values:
- starting capital
- % risked per trade
- penetration factor (1 means the buy trigger is 50-SMA + ATR(10), 2 means 50-SMA + 2ATR(10) and so on).

Not earth shattering results but not bad either. Imagine what this strategy could return if diversified across a range of non-correlated futures markets. Also remember that this is an extremely simple version. For example, it doesn't add positions when a trade works. You can also do a lot of work on portfolio allocation.

This just proves (to me anyway) that one can still get decent return by following simple (and apparently obsolete) strategies, if EOD trading is your game. One can easily programme this on Tradestation and leave it on autopilot.
 

Attachments

  • SMH.zip
    192.2 KB · Views: 794
Last edited:
Running the same system on the data of SPY (S&P500 ETF), over 13 years the return is just over 300%. The drawdowns are not vicious either.

Again, not spectacular but I'm impressed, given the simplicity of the system and the potential.
 

Attachments

  • SPY.zip
    486.2 KB · Views: 562
OpenMind said:
Running the same system on the data of SPY (S&P500 ETF), over 13 years the return is just over 300%. The drawdowns are not vicious either.

Again, not spectacular but I'm impressed, given the simplicity of the system and the potential.

Interesting, thanks for posting.

Some thoughts. I can't see that you factored in execution costs (brokerage + slippage). That might drastically affect the end result. Secondly, and I may be wrong here, I think ETFs are a poor choice of investment vehicle to test a trend following strategy on. Firstly, expensive / difficult to short. Secondly, if you are long, they are wasting assets. i.e. ceteris paribus, they depreciate over time due to management fees.
 
You are right. I didn't factor in costs.

Ideally I'd have liked to test it on historical futures data (which I will do at some point soon), but I don't quite follow you logic against ETFs. You can short ETFs through CFDs. There is no management fee, they are traded just like individual shares. The price data of SPY should be fairly close to the continuous contract data of ES or SP.
 
OpenMind said:
:

Go long at the market if the price closes above 50-SMA + ATR(10)
Go short at the market if the price closes below 50-SMA - ATR(10)
Trail the stop on the 50-SMA.

.

Was hoping someone would ask but I guess I am the dunce around here! :eek:

What is ATR?

Split
 
Splitlink said:
Was hoping someone would ask but I guess I am the dunce around here! :eek:

What is ATR?

Split

Average Trading/True Range?
 
Last edited:
OpenMind said:
Running the same system on the data of SPY (S&P500 ETF), over 13 years the return is just over 300%. The drawdowns are not vicious either.

Again, not spectacular but I'm impressed, given the simplicity of the system and the potential.


Not bad, but well below what I want from a simple system.

"The Turtles became the most famous experiment in trading history because over the next four years, they earned an average annual compound rate of return of 80%."

http://www.originalturtles.org/system.htm
 
new_trader said:
Not bad, but well below what I want from a simple system.

"The Turtles became the most famous experiment in trading history because over the next four years, they earned an average annual compound rate of return of 80%."

The turtle system is simple, but they spent years devising it. It now looks easy, but it was way ahead of its time.

The point I am trying to make here is this: if a simple system which is freely available everywhere and which can be coded and tested in Excel within 2 hours still generates a healthy (15% and 23% in the two test cases, before costs) return, think what one can achieve with a little more research and testing.
 
new_trader said:
Not bad, but well below what I want from a simple system... an average annual compound rate of return of 80%.
Go for it new_trader.

Making 80% p.a. is a piece of p********************. Let us all know when you have bought your Carribean island.
 
LevII said:
Go for it .

Making a piece of Carribean island is p********************. Let us all know when you have bought 80% new_trader.

Its great fun rearranging quotes
 
OpenMind said:
The turtle system is simple, but they spent years devising it. It now looks easy, but it was way ahead of its time.

The point I am trying to make here is this: if a simple system which is freely available everywhere and which can be coded and tested in Excel within 2 hours still generates a healthy (15% and 23% in the two test cases, before costs) return, think what one can achieve with a little more research and testing.

My point is that there are freely available systems that produce much better results. Certainly the ones I have tested do.
 
Testing times

new_trader said:
My point is that there are freely available systems that produce much better results. Certainly the ones I have tested do.

Try testing your theories in real time with real money and let us know how you get on.
 
OpenMind said:
The turtle system is simple, but they spent years devising it. It now looks easy, but it was way ahead of its time.

The point I am trying to make here is this: if a simple system which is freely available everywhere and which can be coded and tested in Excel within 2 hours still generates a healthy (15% and 23% in the two test cases, before costs) return, think what one can achieve with a little more research and testing.

the turtle system, as I understand it, was the Donchian Channel breakout, but with an eye to ATR.
http://www.seykota.com/tribe/Resources/index.htm

You can see something similar on Ed Seykotas website. Where they also test a Simple Moving Average system, and use ATR to determine stop-losses, rather than a fixed amount, thus accounting for volatility.
http://www.seykota.com/tribe/TSP/index.htm

if you read the originalturtles website, you will see they place small trades when the breakout occurs, and add more as the trade goes in their favour, rather than placing the full amount on a signal. thus, when you get the inevitable false signals, you are not always risking the full amount, and can potentially escape from a false signal with losing 1/5 or 1/4 of normal trading size.

these little nuances need to be appreciated.

re: freely available.
Richard Dennis did say that he could post his full trading system for FREE in a newspaper, and people will lose, since one of the key components is sticking with it during drawdown, waiting for the eventual trend to emerge. A number of turtles went belly-up because they starting tweaking the rules after a string of losses.

good luck with your project.
 
new_trader said:
Bookmark this thread and come back in 5 years time :rolleyes:

I was not being sarcastic - consider starting a live journal so we can see yourv idea in action, realtime, with real money. Back testing is worth squat on its own. I've been trading longer than 5 years and used forward live testing until I found something that works. :rolleyes:
 
Level 2

LevII said:
Go for it new_trader.

Making 80% p.a. is a piece of p********************. Let us all know when you have bought your Carribean island.

These newbies have yet to realise that if it was so easy everybody would be using this "system"
He ignores any advice and posts sarcastic retorts. This is why regular traders here now post so little. Why offer advice just to receive saracstic retorts from knowitall newbie ****s. :cheesy:
 
I would not attempt to use the original Turtle system today, you will get your ar*e handed to you.
Simple strats are the best but important to test them thoroughly. I have not looked at the one mentioned at the beginning of the thread but looking at the rules I think that it will still get badly whipped in sideways markets. That is where you should concentrate your efforts. Easy to get on a trend with a strat like this but hard to hold onto money when time series are not strongly directional.
 
trendie said:
re: freely available.
Richard Dennis did say that he could post his full trading system for FREE in a newspaper, and people will lose, since one of the key components is sticking with it during drawdown, waiting for the eventual trend to emerge. A number of turtles went belly-up because they starting tweaking the rules after a string of losses.

good luck with your project.


There is a pdf of the system on their website and Richard says:

"I always say that you could publish my trading rules in the newspaper and no one would follow them. The key is consistency and discipline. Almost anybody can make up a list of rules that are 80% as good as what we taught people. What they couldn't do is give them the confidence to stick to those rules even when things are going bad"

Here are some interesting articles:

http://www.traderslog.com/capturing-trend-days.htm

http://www.traderslog.com/trading-systems.htm

(I think some of them are in the knowledge lab)
 
If you're looking for an excellent "free" long term trend-following system for using across a diversified portfolio of futures markets then you might try Mark Johnson's PGO (pretty good oscillator). He has even conveniently coded it into excel. It's probably not suitable for stock trading as it goes short as well as long and long term trend-following systems do not work so well on the short side of stocks.

www.mjohnson.com

RTN
 
Top