Systems exposed

[barjon]

from Murphy:

......of 12 systems tested from 1975-84 only 4 generated significant profits. Of these four, 2 were channel breakout systems and one was a dual moving average crossover system. A later study published in 1990 compared 23 technical trading systems on data from 1976 -86. Once again channel breakout systems and moving average systems came out on top. It was concluded that a channel breakout system was the best starting point for all technical trading system development

The final conclusion of all this research was that the most successful of all systems tested was Donchian's 4 week rule which is simplicity itself:

1. Cover short positions and buy long whenever the price exceeds the highs of the four preceding full calendar weeks.

2. Liquidate long positions and sell short whenever the price falls below the lows of the four preceding full calendar weeks.

Wonder if anything's changed

good trading

jon

 

[trendie]

Hi barjon,

I pointed out the exact same passage on the "Your One Line Trading Tip" thread, post number 130.

That was the one passage which underlines the fact that no-one can consistently predict the market, hence trend-following, MAs etc are so logical to use.

It is also brain-numbingly simple to use.

However, the drawdowns test your resolve !!

 

[Pippppin]

Trendie

I am often suprised at the number of people on T2W who make statements of fact based on their own experience/perception

Maybe your post should read:-
"no-one i am aware of can consistently predict the market"

 

[trendie]

Trendie

I am often suprised at the number of people on T2W who make statements of fact based on their own experience/perception

Maybe your post should read:-
"no-one i am aware of can consistently predict the market"

I stand corrected.

 

[bracke]

barjon

An interesting post

When talking of lows and highs are they intraday or end of day ?

Have you used the system ?

trendie

Have you used the system, if so on what have you used it, did it work for you and what drawdown did you get ?

Regards

bracke

 

[trendie]

barjon

An interesting post

When talking of lows and highs are they intraday or end of day ?

Have you used the system ?

trendie

Have you used the system, if so on what have you used it, did it work for you and what drawdown did you get ?

Regards

bracke

paper-traded - Dow and SP500.
I also preferred the shorter 10-day channel breakout as a get-out to protect profits.
As you know, the markets do not trend or range consistently, ( as far as I am aware ), so I lean towards EMA and SMA to react to the changing volatility, which I prefer to use.

I just need to learn to follow my own rules, and I am laughing !!!

 

[barjon]

trendie - sorry for stealing your thunder

bracke - no idea

 

[trendie]

trendie - sorry for stealing your thunder

bracke - no idea

not at all, barjon.
glad another trader has seen the absurdity of the claims of system-providers.

Although the tests were conducted a decade or two ago, I am sure the principles are still the same.

The really difficult thing is to accept that super-computer giga-hertz processing can be beaten by a line drawn on a chart with a crayon !!

 

[TWI]

I think the secret with these sorts of simple breakout systems is diversification of market. The more markets you are able to include the smoother your equity curve. I did a lot of work on a simple breakout commodity trading system a few years ago for a CTA covering 23 commodity futures markets. Any single market had an unacceptable Sharpe ratio but the combination was really quite good. Naturally another important consideration here is the weighting given to each market.
I think diversification of markets with a system and/or systems in a market is essential if you want to be consistent in your returns.

 

[bracke]

Looking at the ftse revealed the following:

end of day high on 21/09 @ 4609 and intraday high on 22/09 @ 4631
These exceeded the highs of the previous 4 weeks.

If I am understanding the system correctly I should have bought the index following these highs. If I had where would I have placed the stop ?

Whilst I am not a timid person my bravery does not extend to such a purchase neither does my bank balance.

All of the above 'as far as I am aware'

Regards

bracke

 

[oatman]

2. Liquidate long positions and sell short whenever the price falls below the lows of the four preceding full calendar weeks.
Is that what you mean?

 

[FetteredChinos]

these are essentially the turtles rules arent they, give or take a few day's channelling..

arent the last 5 or so years in drawdown? to the extent that there are various fading strategies out there? eg Turtle Soup for one..

ie fade 10 day breakouts, and wait for a return to the 20 day MA is one that i seem to recall working from a few backtests ago....

had about a 75% hit rate if i recall.

FC

 

[barjon]

Looking at the ftse revealed the following:

end of day high on 21/09 @ 4609 and intraday high on 22/09 @ 4631
These exceeded the highs of the previous 4 weeks.

If I am understanding the system correctly I should have bought the index following these highs. If I had where would I have placed the stop ?

Whilst I am not a timid person my bravery does not extend to such a purchase neither does my bank balance.

All of the above 'as far as I am aware'

Regards

bracke

bracke

You wouldn't have a stop as such. After your long entry you would close it and go short when the price fell below the low of the previous 20 days. That level might seem a very long way away to start with but it's progressive as each 21st day drops off (if you see what I mean!!)

 

[bracke]

bracke

You wouldn't have a stop as such. After your long entry you would close it and go short when the price fell below the low of the previous 20 days. That level might seem a very long way away to start with but it's progressive as each 21st day drops off (if you see what I mean!!)

barjon

Thank you for your reply

Referring to my example eod 21/09 @ 4608

The eod low of the previous 20 days was 23/08 @ 4405 which means a 203 point drop before closing the trade and is a little too rich for me.

Since the 21/09 the lowest close was 27/09 @ 4541 a 67 point drawdown.

The index has not closed above 4608.

I would not be happy with this trade.

Not quite clear on the 21st day dropping off.

Regards

bracke

 

[Roberto]

The really difficult thing is to accept that super-computer giga-hertz processing can be beaten by a line drawn on a chart with a crayon !!

Not if you're a bit of a technophobe like me ... it's really _most_ gratifying!

 

[fastnet]

Fettered Chinos - sure it's a turtle in another guise. I have nattered on about systems to trade big trends on these boards before. These days - like you - I fade such breakouts.

Potted history is that I started trading an turtle-type system on an EOD basis but got whipsawed to death. Also the idea of buying right at the recent high really didn'y appeal.

I now work on the principle that such highs will almost always sink back to re-test the o/h resistance they defeated whether or not they then sink back into the channel or break free. Of course now and then they race away on a new trend and don't look back. These occasions are my drawdowns - and there's no particular place to stop out. I hope to modify the system so that I can stop and reverse these occasional runaway new trends.

Also see discussions by member called ''aussietrader'' who I know trades turtle type set-ups.

 

[TWI]

I tried this 4 week break and the one with the MA. All the markets I tried it on showed pretty appaling results even though many ended net up, the volatility of returns was totally unacceptable and would definitely put one into an early grave.
Have not tried to combine many markets together but even if I did the results at best would be a few % per annum, certainly not double figures. Maybe it works over 20 years but for a 3-5 year plan I do not recommend it.

 

[Roberto]

...
... the volatility of returns was totally unacceptable and would definitely put one into an early grave ...
...
Maybe it works over 20 years but for a 3-5 year plan I do not recommend it.

I think this is exactly right. If you think about the "original turtles" (there were ... what ... 14 of them?) they were between them covering almost every available market, which must overall smooth things out tremendously, surely? Very different for a "sole trader", I think ...

One of the (many) reasons I prefer intraday trading, myself, is that by entering and exiting a comparatively large number of comparatively small-stake trades, you get some of this "smoothing out effect" that way as well. Sure, you can have some losing runs, but no frightening volatility of results leading to an early grave.

 

[stevespray]

I did some work on some price channel breakout systems a year or so back. At first glance some of the results seemed pretty random. However, a closer look revealed some more interesting facts. Certain results improve significantly when you apply certain filters. If my memory serves me correctly, the turtles system, which was a breakout system, used a filter which tested to see if the last suggested trade was good or bad. I think they decided that they would only take the signal if the last trade had failed.

The key here is back testing. My personal opinions are based mainly on back testing of S&P500 Futures. Seeing as it has 500 constituent parts and is the most liquid market available, my thinking is that it is the most difficult to tame. Therefore, in my opinion, any system which does okay on S&P’s is likely to be adaptable to other markets.

This thread has given me a taste for further research.

How about the a 20 day price channel breakout if the last signalled trade was in the same direction and made money ?
For example, our system signals a long. We would check back to see if the last signal would have yielded a profit if we would have taken the signal.

I’ll have a fiddle on Trade Station.

Steve.

 

[Roberto]

If my memory serves me correctly, the turtles system, which was a breakout system, used a filter which tested to see if the last suggested trade was good or bad. I think they decided that they would only take the signal if the last trade had failed.

The rules were that they did that for the 20-day break-out system, yes (but not for the 55-day break-out system).