|
PureBytes Links
Trading Reference Links
|
Hello Steve,
In my very humble opinion you should optimize for expectation - high
expectation will take care of the drawdowns and at the same time will
make the performance consistent.
Just make the optimization as you normally do it in the AA. Then
export the results to Excel. In the output columns you have enough
information to calculate the expectancy
(Avg.Win*PercentProfitable/Avg.Loss - PercentLosing)
and add it to a new column. Then use the 3D analyser to see plateaux
for expectancy.
Even better, you could take a look at the expectancy
figure, multiplied by the number of deals.
Regards,
Alexander Dimitrov
SD> Rick,
SD> Hermans 3D optimizer is a blast to play with. I am having great fun
SD> exploring new ideas. You make a good point, sharp peaks bad, rolling hills
SD> good. I will look for that.
SD> I'm not sure if I'm adequately expressing my original point. The z-axis of
SD> the 3D plot is total system return over the 50-year test. I am suggesting
SD> perhaps there is some other system measurement that could be used for the
SD> z-axis. I need a mathematical expression that evaluates the return of the
SD> system but also takes into account the volatility of the equity curve.
SD> Thanks for the discussion. Putting this into words helps me clarify my own
SD> thoughts.
SD> -Steve
SD> -----Original Message-----
SD> From: Rick Parsons [mailto:RickParsons@x...]
SD> Sent: Thursday, October 17, 2002 2:23 PM
SD> To: amibroker@xxxxxxxxxxxxxxx
SD> Subject: RE: [amibroker] Optimizing for consistent performance
SD> Steve,
SD> Re: >>it is designed to optimize the parameters to maximize the total
SD> system return.<<
SD> Just my humble opinion but I suspect you are using the 3-D for the wrong
SD> purposes. If you want maximum returns, you can simply optimize in AA and
SD> sort by % return.
SD> 3-D is to show you which parameters are the most robust. If you find a
SD> large flat plain in the 3-D that means there is a wide range of parameters
SD> that give consistent returns. This is robustness.
SD> If you use 3-D to find the best return, chances are it is located on a
SD> peak with sharp drop-offs (although this is not always the case). The sharp
SD> drop-offs indicate that the parameters are not robust.
SD> To some extent, robustness helps with "consistent performance".
SD> Rick
SD> -----Original Message-----
SD> From: Steve Davis [mailto:sdavis@x...]
SD> Sent: Thursday, October 17, 2002 2:07 PM
SD> To: amibroker@xxxxxxxxxxxxxxx
SD> Subject: RE: [amibroker] Optimizing for consistent performance
SD> Rick,
SD> Yes, I have tried Herman's optimizer -- it is designed to optimze the
SD> parameters to maximize the total system return. That is useful, but not
SD> exactly what I want. If I optimize the system to maximize gains, the
SD> drawdowns become horrible. I need a system that produces steady monthly
SD> returns.
SD> You can get at least 50 years of historical EOD data from Yahoo. The
SD> DOW, SP500, Utilities, and transports all have at least 50 years of data
SD> available.
SD> -Steve
SD> -----Original Message-----
SD> From: Rick Parsons [mailto:RickParsons@x...]
SD> Sent: Thursday, October 17, 2002 2:01 PM
SD> To: amibroker@xxxxxxxxxxxxxxx
SD> Subject: RE: [amibroker] Optimizing for consistent performance
SD> Steve,
SD> Not exactly sure what you are asking. But have you tried Herman's 3-D
SD> Optimizer for Excel (in download area, I believe) to find the most robust
SD> parameters?
SD> Another question: Where did you get 50 years of data?
SD> Thanks,
SD> Rick
SD> -----Original Message-----
SD> From: Steve Davis [mailto:sdavis@x...]
SD> Sent: Thursday, October 17, 2002 1:34 PM
SD> To: amibroker@xxxxxxxxxxxxxxx
SD> Subject: [amibroker] Optimizing for consistent performance
SD> AB rocks! After using AB for only a month, I have developed a system
SD> that produces 50% risk-adjusted-returns trading the S&P500 over the last 50
SD> years. It is a short-term trading system with an average trade lasting 4
SD> days.
SD> But there is some spikiness in the equity curve. I am having some
SD> difficulty tuning the system to produce steady monthly cash flow. The goal
SD> is not to maximize profits over 50 years, but to produce consistent
SD> profitable results.
SD> One problem is how to measure the system performance so it can be
SD> properly tuned. I do not know many important statistics of the system. For
SD> example, has the system ever had a losing year? What is the worst-case
SD> number of consecutive losing months? I don't think minimizing drawdown is
SD> sufficient to meet my goals.
SD> The approach I have been using is an iterative process of
SD> eye-balling the equity curve, changing parameters, and trying again. There
SD> must be a better way. Is there a reasonable way to automate this task?
SD> How are other AB users optimizing their equity curve?
SD> Cheers,
SD> -Steve
SD> Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
SD> (Web page: http://groups.yahoo.com/group/amiquote/messages/)
SD> Check group FAQ at:
SD> http://groups.yahoo.com/group/amibroker/files/groupfaq.html
SD> Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
SD> Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
SD> (Web page: http://groups.yahoo.com/group/amiquote/messages/)
SD> Check group FAQ at:
SD> http://groups.yahoo.com/group/amibroker/files/groupfaq.html
SD> Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
SD> Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
SD> (Web page: http://groups.yahoo.com/group/amiquote/messages/)
SD> Check group FAQ at:
SD> http://groups.yahoo.com/group/amibroker/files/groupfaq.html
SD> Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
SD> Yahoo! Groups Sponsor
SD> ADVERTISEMENT
SD> Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
SD> (Web page: http://groups.yahoo.com/group/amiquote/messages/)
|