Breakout systems

[Sean K]

Yep, I agree totally with the trade each time frame.

That's pretty random though isn't it?

Aren't you trading mechanically?

Sorry if I've missed some other factors...

 

[tech/a]

Aren't you trading mechanically?

Not since July last year.
But still in the context of Trading Breakout systems These are elements discussed above I would be considering.

 

[julius]

Could someone who is trading breakouts mechanically post a couple of entry parameters... just to clarify 'breakout'

I've always found mech. breakout systems test poorly.

 

[nizar]

It is my belief that a breakout as an entry is a very poor entry per se when assessed over all market conditions, significantly underperforming random entry in profitability and drawdown.

Why havent you tested this belief?
I would be interested to see some results on this.

I believe -- and I could be wrong, that a breakout entry and a nice exit will outperform a random entry with the sane exit. Maybe not significantly -- But I would be very suprised to see a significant outperformance by the random system.

I believe this is partly because it is known to be a very popular entry trigger and partly because there is a strong tendency towards mean reversion after a breakout.

There is a strong tendancy towards mean reversion after the breakout I agree -- BUT this only occurs in the very short term (i have tested this).

If your stop/exit is wide enough to allow for this, and your average trade holding time (trade length is defined primarily by the exit) is several weeks, then the breakout system will outperform I would think.

Note that the above statements are NOT mutually exclusive of a statement that breakout systems can be very profitable - it's just that they can be profitable DESPITE the entry, not BECAUSE of it.

Yes I agree. Doesn't mean a random entry would outperform it though.

 

[nizar]

Could someone who is trading breakouts mechanically post a couple of entry parameters... just to clarify 'breakout'

I've always found mech. breakout systems test poorly.

H>Ref(HHV(H,X),-1)

In english:
The last bar High is greater than the Highest High of the last X bars.

 

[Nick Radge]

The classic breakout model is the Donchian method. Basically if prices are above a 200-ma and then breakout above the highest high in n days, where n can be any length. The trade is closed when prices break back below x days, where x is shorter than n.

In general terms, a high n will create less trades and keep the system out of much noise. As a rule of thumb, the higher x is the larger the trend that will be captured, although the larger the open profit giveback is.

In modern day breakout systems, traders attempt to add other filters some of which are macro based. As an example, breakout systems will be working well in Coal, Oil and Energy sectors on the upside and perhaps Financials and Discretionary Spending on the downside.

Many classic breakout models are used in commodities rather than equities for correlation reasons.

I traded the original Turtle model, a variant of Donchian, back in 2001 for about 8-months and found it quite tough.

This post may contain advice that has been prepared by Reef Capital Coaching ABN 24 092 309 978 (“RCC”) and is general advice and does not take account of your objectives, financial situation or needs. Before acting on this general advice you should therefore consider the appropriateness of the advice having regard to your situation. We recommend you obtain financial, legal and taxation advice before making any financial investment decision.

 

[theasxgorilla]

The reason I ask is because I originally developed a breakout system based on 2005 - 2007 data. It worked well when ran live late 07, but now I've put data from 97 - 2004 into the system and results are fairly poor.

Which market? What was your universe?

 

[MichaelD]

Why havent you tested this belief?
I would be interested to see some results on this.

I *have* tested this belief quite extensively. Compared with random entry, ALL conventional methods of trading breakouts perform worse, particularly in the area of drawdown.

A good starting point for a breakout is the one built into Metastock;

ROC(CLOSE,1,percent) >= 5 AND VOLUME >= (Mov(VOLUME,50,EXPONENTIAL)*1.5)

Again, however, I am NOT saying that using a breakout as an entry signal is unprofitable, 'cause it is, just that it is profitable because of OTHER factors, not the entry per se.

Breakouts fascinate me.

 

[theasxgorilla]

Could someone who is trading breakouts mechanically post a couple of entry parameters... just to clarify 'breakout'

I've always found mech. breakout systems test poorly.

How about this one?

Entry setup:
* Today's high is greater than the 40-day exponential moving average (EMA) of closes AND
* Today's high is the highest high for the last 70 days AND
* Today's close must be higher than today's open.

Entry trigger:
* Buy at tomorrow's open when today's high crosses the highest high of the last 10 days.

Give that a whirl. Be sure to try it between 1st of Jan 07 and today...

 

[theasxgorilla]

Again, however, I am NOT saying that using a breakout as an entry signal is unprofitable, 'cause it is, just that it is profitable because of OTHER factors, not the entry per se.

I think it depends on how you 'frame' your entry. The effectiveness of a breakout, and it's ability to overcome mean reversion, can be increased if you frame the setup conditions adequately and prevent the trigger from biting at ever breakout that swims by.

 

[Nick Radge]

ALL conventional methods of trading breakouts perform worse

What is a conventional method?

 

[weird]

Hi Synergy, here is some historical backtested results over 10 years of a donchian based breakout system I traded during 2005-2008 on ASX stocks to compare.

The actual results mirror the backtested results quite closely. It uses a few filters to trigger the breakout.

The top of the equity curve is around the 1/7/2007.

This actual system does not play a large part in my actual trading, actually 0 atm (the system has pretty much switched itself off over the last 9 months), however the results perhaps may beat some instances of simple buy and hold.

 

[tech/a]

How about this one?

Entry setup:
* Today's high is greater than the 40-day exponential moving average (EMA) of closes AND
* Today's high is the highest high for the last 70 days AND
* Today's close must be higher than today's open.

Entry trigger:
* Buy at tomorrow's open when today's high crosses the highest high of the last 10 days.

Give that a whirl. Be sure to try it between 1st of Jan 07 and today...

Looks familiar wonder where you got that one from?

 

[Boggo]

Tech, your recent acquisition NHC on "my" breakout system.

Mike

.

 

[MichaelD]

What is a conventional method?

Anything that the majority of traders are aware of and believe to be profitable. When one trades with the crowd, one will get crowd results.

 

[stevo]

Could someone who is trading breakouts mechanically post a couple of entry parameters... just to clarify 'breakout'

I've always found mech. breakout systems test poorly.

I think that some will benefit from such beliefs

It's a bit like saying that you don't like vegetables. There must be as many breakout systems as vegetables to try, probably more. There must be some out there that are worth eating.

Does it really matter what we label a system - breakout, Overbought / Oversold, contrarian etc - as long as it works.

In terms of random entries and exits have a look at Random entries & Exits and Random Entry on my blog.

 

[julius]

stevo,

I've got no clue what your on about.

Vegetables?

 

[tech/a]

Julius

Here is one of mine
Techtrader.
Its a long timeframe long only breakout system.
Its traded live here and has been for 6 yrs.
There are hundreds of pages of info including
coding and Q&A.

http://www.thechartist.com.au/forum/ubbthreads.php?ubb=postlist&Board=53&page=1

 

[nizar]

stevo,

I've got no clue what your on about.

Vegetables?

Julius.

His point is that just because you havent been able to find a profitable breakout system, doesnt mean there arent any out there.

 

[julius]

I just wanted to clarify the definition of a 'break-out' system...apples with apples, etc

To be honest, I haven't spent much time testing long term systems on equities.

I'm more inclined toward higher frequency models - just a personal preference.