[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [amibroker] Swing Trading - Strategy Performance



PureBytes Links

Trading Reference Links

Paul --
Thank you very much for your detailed response.  This is much more than I expected.  But I do appreciate it all.

Wow.
-- Keith


Paul Ho wrote:
As you probably know, it was Edwards and McGee that poparlised chart patterns like head and shoulders, triangles, double bottoms. But they are usually subjective and doesnt lend itself well in conversion to mechanical systems. But some modern, and sometimes less well know patterns have defintions that can be coded. One such example is the NR4 pattern defined in the Nqoos.com site. Clyde Lee also defined a number of swing patterns and listed on his web site.  When you code them, you'll find that you wont need any oscillators or any cyclical indicators as a general rule. Zig Zag can be used to defined patterns but itself is not a pattern. If you do use Zig Zag, you got to be aware that it looks into the future. so you 're only safe to use it to mark the second to last swing point. For example, if you're long, you can only use zig zag to find the last high, but not the last low. Take a look at those sites and see if that becomes clearer for you.
I optimize my systems over the period of 3.5 years. I then backward test for another 8 years and forward test it another 9 months. I test my sytems over about a thousand symbols and have generated thousands of signal and at least several hundred signals ended up in trades. I didnt paper trade my system initially. but I examine a hundred of these trade visually over time before I go pilot. In pilot, I use only some of the signals and I checked every trade visually before committing myself. When I was happy, I went live.  Today I'm still refining it using real trading results as feedback. I also find it extremely useful to have a trading buddy, not one that you necessarily share you systems with. but one that you can use to bounce  your ideas around and discuss similar problems or observation that you encounter in your development cycle.
 

Paul


From: amibroker@xxxxxxxxxxxxxxx [mailto:amibroker@xxxxxxxxxxxxxxx] On Behalf Of Keith McCombs
Sent: Wednesday, 8 February 2006 7:38 PM
To: amibroker@xxxxxxxxxxxxxxx
Subject: Re: [amibroker] Swing Trading - Strategy Performance

Paul --
I found your post very interesting.

Could you expand a bit on "pattern based systems".  Are you referring to patterns such as zig-zag and candlestick formations or to something else.  I am not asking what you use but am instead trying to understand your terminology.

Also, over what period of time, and how broad a data base, did you run your backtest before deciding to trade it for real?  How about walk forward testing?

Thank you very much for your input.
-- Keith

Paul Ho wrote:
Ara,
I have been using my swing system for about 3 months . I have tested thousands of variations.
1. the latests backtest
has a RAR of 110% and CAR of 67% Avg Proit/loss of 12.4%, 61.5% winners Max Drawdown of 5.6% and expectancy of 40cents per 100$ invested per day (this is my own definition) and profit factor of 6.64
2. Real performance is also very good with figures matching backtest in terms of win/loss ratio, as well as Avg Profit (Avg profit for each trade). Max drawndown is usually worse at around 8%. I dont believe 5.6% is sustainable (you got to have a bad day sometime).
3. Position Sizing is volatility based with a maximum cap of 5.7% of my initial capital.  Volatility risk is at 0.6% of my capital. Because I trade of a lot of smallcap stocks. I also pay great attention to turnovers and capitalisations and have included them  in my position sizing. I dont compound my profit in my testing all positions are sized based on initial capital. I have tested it with compounding for interest but I think it is distorting to select systems based on those figures.
4. I find that pattern based systems usually gives me better results than purely indicator based systems. In pattern based system I mean systems that evaluates the relative position of price bars to form pattern. I find that volume is very useful but it was initially difficult to incorporate it into my system but once I have successfully incorporated it. I really turbo charge my system.  Generally I find swing systems are very good pattern system. but if you use standard indicators to code your system. I find that it will stop working all together at some point. This is when the market has changed it rhythm and both tempo and the magnitude of the swing has changed. Clyde Lee's ideas about swing is very good and its well worth a read on his site. Unfortuntely he only uses tradestation.
I dont scale in or out, stop is at 16%. scaling in and out might be the next thing to consider.
 
 

Paul



From: amibroker@xxxxxxxxxxxxxxx [mailto:amibroker@xxxxxxxxxxxxxxx] On Behalf Of Ara Kaloustian
Sent: Monday, 6 February 2006 6:34 AM
To: AB-Main
Subject: [amibroker] Swing Trading - Strategy Performance

Would like to get a poll of sorts .... for expected swing trading performance. Maybe there is some information that can help everyone regardless of specific system used.
 
Primarily interested in the potential of a system, any pitfalls / difficulties in implementing it after simulation studies were completed...
 
What would make good statistical info:
 
1. Backtest performance
2. Real performance / Any special considerations in implementing
3. Money management - scaling in/out, stops, risk level...
4. Any issues and concerns
 
My intent / desire is NOT to delve into anyone's specific system as such, so if you feel it is appropriate to describe your system, please keep it at a high level.
 
I have toyed with several systems, including Relative strength, combination of standard indicators (MACD, Stochastics and CCI) with preliinary backtest results of 20% per year, +/- a little, but have not really traded based on any system as such, certainly not with any consistancy.... so no track record.
 
I have not succesfully integrated market environment in backtests yet. My next step.
 
My hope is that I can make about 5% to 10% per month using trades of 3 days to 2 weeks. I have been known to be too optimistic ... at times ...
 
Appreciate any feedback
 
Ara
 


Please note that this group is for discussion between users only.

To get support from AmiBroker please send an e-mail directly to
SUPPORT {at} amibroker.com

For other support material please check also:
http://www.amibroker.com/support.html





SPONSORED LINKS
Investment management software Real estate investment software Investment property software
Software support Real estate investment analysis software Investment software


YAHOO! GROUPS LINKS