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

RE: [amibroker] Re: System Performance Indicators [was: Can someone fix this OLE code?]



PureBytes Links

Trading Reference Links

Your post is contradictory of positions you have stated in the past and even with other parts of your post …

If we define the FrontEnd as where we trade (AB as trading platform,
AT or as a cue for manual entry of trades) and the BackEnd as
SystemDesign&Evaluation then, so far I haven't found a use for
PerformanceMetrics at the FrontEnd.”

 

Doesn’t the above essentially state that Performance Metrics have no value in trading ? 

 

If that’s the case what do you do when your OOS MDD exceeds the previously tested systems MDD ? 

 

Pretend it didn’t happen ?

 

Here again you state the same thing in a different way …

 

If SD&E is the production line then the trades I take are the product.
I don't see any benefit in adjusting the production line based on
what I observe from a short production run.

 

Trading is by definition Out Of Sample.  Performance metrics are what make up the objective function.  

 

Isn’t the implication of your statement that there is no value to the out of sample performance metrics and doesn’t that then imply that there is no value to the “objective function”.  

 

How did I misread what you stated ?

 


From: amibroker@xxxxxxxxxxxxxxx [mailto:amibroker@xxxxxxxxxxxxxxx] On Behalf Of brian_z111
Sent: Wednesday, May 21, 2008 11:35 PM
To: amibroker@xxxxxxxxxxxxxxx
Subject: [amibroker] Re: System Performance Indicators [was: Can someone fix this OLE code?]

 

> If you don't see any value in adjusting the "production line" based
>on the
> performance metrics then that would suggest you've not yet traded
>with
> dollar 1 . To use your "production line" example, Quality Assurance
>is about
> seeing what comes off the production line and adjusting the line
>when
> tolerances are out of whack . This due diligence is used in
>virtually every
> manufacturing process and should be used in trading as well.

Yes, that is what I said.
You are mis-reading me.

> Money Management is a function of the SYSTEM it is not a
>PERFORMANCE METRIC
> nor is it an OBJECTIVE FUNCTION which would typically be used on
>OOS data to
> evaluate the systems performance whether that be in test or with
>real money.
> If you don't believe in adjusting the "production line" based on
>performance
> metrics then you also don't believe in objective functions . How
>could you ?
>

I am talking about a Portfolio Metric.

I can't be bothered explaining that any further.

> Why is it when I read your posts I feel like I'm talking to a 12
>year old
> who's stuffing nickels and dimes in a piggy back waiting for the
>day when he
> can crack it open to finance his first "adult experience" .

I can't account for your reading list, your comprehension or your
feelings Fred.
You will have to do that for yourself.

brian_z

--- In amibroker@xxxxxxxxxps.com, Fred Tonetti <ftonetti@xx.> wrote:
>
> If you don't see any value in adjusting the "production line" based
on the
> performance metrics then that would suggest you've not yet traded
with
> dollar 1 . To use your "production line" example, Quality Assurance
is about
> seeing what comes off the production line and adjusting the line
when
> tolerances are out of whack . This due diligence is used in
virtually every
> manufacturing process and should be used in trading as well.
>
>
>
> Money Management is a function of the SYSTEM it is not a
PERFORMANCE METRIC
> nor is it an OBJECTIVE FUNCTION which would typically be used on
OOS data to
> evaluate the systems performance whether that be in test or with
real money.
> If you don't believe in adjusting the "production line" based on
performance
> metrics then you also don't believe in objective functions . How
could you ?
>
>
>
> Why is it when I read your posts I feel like I'm talking to a 12
year old
> who's stuffing nickels and dimes in a piggy back waiting for the
day when he
> can crack it open to finance his first "adult experience" .
>
>
>
> _____
>
> From: amibroker@xxxxxxxxxps.com [mailto:amibroker@xxxxxxxxxps.com]
On Behalf
> Of brian_z111
> Sent: Wednesday, May 21, 2008 7:55 PM
> To: amibroker@xxxxxxxxxps.com
> Subject: [amibroker] Re: System Performance Indicators [was: Can
someone fix
> this OLE code?]
>
> If we define the FrontEnd as where we trade (AB as trading
platform,
> AT or as a cue for manual entry of trades) and the BackEnd as
> SystemDesign&Evaluation then, so far I haven't found a use for
> PerformanceMetrics at the FrontEnd.
>
> To me the BackEnd is where we analyse a sample of trades (the trade
> series) as part of the design process (just as you have described
it).
>
> I have also plotted performance metrics, as visual learning aids (I
> scroll from stock to stock comparing how my system ideas look in
> different symbols/instruments/assests).
>
> I also create metric reports, via explorations, for a W/L of stocks
> etc for similar reasons.
>
> (The methods that have come out of this thread give me more
efficient
> ways to do that).
>
> In that regard I consider them to be tools that are part of the
SD&E
> process.
>
> If SD&E is the production line then the trades I take are the
product.
> I don't see any benefit in adjusting the production line based on
> what I observe from a short production run.
>
> I think we should be careful as to what we call portfolio trading.
> I don't see taking a few trades, from the total number available in
a
> W/L of stocks, as portfolio trading.
>
> To me Portfolio Trading is where we use MoneyManagement, the
profile
> of the trade and correlation to make our decisions in a synthetic
way.
>
> In simple terms our ObjectiveFunction/FitnessMeasure should include
> MM, correlation and the trade profile, as parameters, and be chosen
> by backtesting and then adjusted after a reasonable number trades
(if
> ongoing adjustment is the users policy).
>
> Once again, IMO, that adjustment should be made in the backend and
> not at the front.
>
> Having said that, I understand the arguments for equity curve
trading.
> I personally reject them but if others succeed with that method
good
> luck to them.
>
> brian_z
>
> --- In amibroker@xxxxxxxxx <mailto:amibroker%40yahoogroups.com>
ps.com, Fred
> Tonetti <ftonetti@> wrote:
> >
> > Personally I only use performance metrics in the following ways .
> >
> >
> >
> > - As feedback to know how my in sample performs i.e. an
> initial
> > milestone
> >
> > - As additional feedback to compare out of sample results
> with in
> > sample and/or to constant yardsticks to get a feel for whether or
> not a
> > system is tradable
> >
> > - As ongoing feedback to determine whether or not it is
> time to
> > reoptimize and/or redevelop
> >
> >
> >
> > I've been down the road you are on but for me it was in essence a
> dead end.
> > I hope you have better success with it.
> >
> >
> >
> > _____
> >
> > From: amibroker@xxxxxxxxx <mailto:amibroker%40yahoogroups.com>
ps.com
> [mailto:amibroker@yahoogrou <mailto:amibroker%40yahoogroups.com>
ps.com]
> On Behalf
> > Of Herman
> > Sent: Thursday, May 22, 2008 6:40 AM
> > To: Fred Tonetti
> > Cc: amibroker@xxxxxxxxx <mailto:amibroker%40yahoogroups.com>
ps.com
> > Subject: Re: [amibroker] Re: System Performance Indicators [was:
> Can someone
> > fix this OLE code?]
> >
> >
> >
> > Thank you very much Fred,
> >
> >
> >
> > When using performance metrics in system design they have (imo)
> more meaning
> > if they should be based on single price/equity arrays. I think
that
> > portfolio metrics should not be used for anything else then to
tell
> you how
> > your portfolio system performs.
> >
> >
> >
> > herman
> >
> >
> >
> > For tips on developing Real-Time Auto-Trading systems visit:
> >
> > http://www.amibroke <http://www.amibroke
> <http://www.amibroker.org/userkb/> r.org/userkb/> r.org/userkb/
> >
> >
> >
> > Wednesday, May 21, 2008, 5:50:54 PM, you wrote:
> >
> >
> >
> >
> > >
> >
> > Herman,
> >
> >
> >
> > I modified the script to allow running Explore's as well as the
> variety of
> > different types of Backtests and Optimizes all based on the
command
> line
> > arguments you supply
> >
> >
> >
> > See the notes at the top of the Script as to how to use and the
> default
> > values.
> >
> >
> >
> > I also renamed the Script to RunAA since that is now more
> appropriate.
> >
> >
> >
> > Individual Backtests appear to run significantly faster than
> Portfolio
> > oriented ones.
> >
> >
> >
> > Fred
> >
> >
> >
> > _____
> >
> > From: amibroker@xxxxxxxxx <mailto:amibroker%40yahoogroups.com>
ps.com
> [mailto:amibroker@yahoogrou <mailto:amibroker%40yahoogroups.com>
ps.com]
> On Behalf
> > Of Herman
> >
> > Sent: Thursday, May 22, 2008 5:09 AM
> >
> > To: Fred
> >
> > Subject: Re: [amibroker] Re: System Performance Indicators [was:
> Can someone
> > fix this OLE code?]
> >
> >
> >
> >
> > Yes, sorry should have been more explicit. The code works nice
and
> there are
> > possibilities to running the code from the tools menu. But this
> would
> > require Explores and Backtests. Not enough time to try all the
new
> ideas
> > :-) This is the code you posted:
> >
> >
> >
> > RunType = "BACKTEST"
> >
> > Times = 1
> >
> > Refresh = 0.25
> >
> > ShowAA = 1
> >
> >
> >
> > Set oArgs = WScript.Arguments
> >
> >
> >
> > For i = 0 to oArgs.Count - 1
> >
> > Arg = Split(oArgs(i), "=")
> >
> > Arg(0) = UCase(Trim(Arg(0)))
> >
> > If Arg(0) = "RUNTYPE" Then
> >
> > RunType = UCase(Arg(1))
> >
> > ElseIf Arg(0) = "TIMES" Then
> >
> > Times = CCur(Arg(1))
> >
> > ElseIf Arg(0) = "REFRESH" Then
> >
> > Refresh = CCur(Arg(1))
> >
> > ElseIf Arg(0) = "SHOWAA" Then
> >
> > ShowAA = CCur(Arg(1))
> >
> > End If
> >
> > Next
> >
> >
> >
> > Set oAB = CreateObject("Broker.Application")
> >
> > Set oAA = oAB.Analysis
> >
> > oAA.ShowWindow(ShowAA)
> >
> >
> >
> > BegTime = Now()
> >
> >
> >
> > While Times >= 0
> >
> > If RunType = "OPTIMIZE" Then
> >
> > oAA.Optimize(0)
> >
> > Else
> >
> > oAA.Backtest(0)
> >
> > End If
> >
> > If Refresh > 0 Then
> >
> > WScript.Sleep Refresh * 1000
> >
> > End If
> >
> > Times = Times - 1
> >
> > Wend
> >
> >
> >
> > EndTime = Now()
> >
> >
> >
> > MsgBox CStr(BegTime) + vbCrLf + CStr(EndTime), 0, "BackTest"
> >
> >
> >
> > ////////////
> >
> >
> >
> > For tips on developing Real-Time Auto-Trading systems visit:
> >
> > http://www.amibroke <http://www.amibroke
> <http://www.amibroker.org/userkb/> r.org/userkb/> r.org/userkb/
> > <http://www.amibroke <http://www.amibroker.org/userkb/>
r.org/userkb/>
> >
> >
> >
> > Wednesday, May 21, 2008, 12:27:50 PM, you wrote:
> >
> >
> >
> > > Herman,
> >
> >
> >
> > > Are you referring to the script I wrote ?
> >
> >
> >
> > > = = = = = = = = = = = = = = =
> >
> >
> >
> > > Hello Fred, can you tell me:
> >
> >
> >
> > > 1) How do i make your code run individual backtests on a
> watchlist? It
> >
> > > default to portfolio BTs.
> >
> >
> >
> > > 2) How do I modify the code to run an explore?
> >
> >
> >
> > > Many thanks!!!
> >
> >
> >
> > > herman
> >
> >
> >
> >
> >
> >
> >
> > > ------------------------------------
> >
> >
> >
> > > 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 NEW RELEASE ANNOUNCEMENTS and other news always check
DEVLOG:
> >
> > <http://www.amibroke <http://www.amibroker.com/devlog/>
r.com/devlog/> >
> http://www.amibroke <http://www.amibroker.com/devlog/> r.com/devlog/
> > <http://www.amibroke <http://www.amibroker.com/devlog/>
r.com/devlog/>
> >
> >
> >
> > > For other support material please check also:
> >
> > <http://www.amibroke <http://www.amibroker.com/support.html>
> r.com/support.html> >
> http://www.amibroke <http://www.amibroker.com/support.>
r.com/support.
> > <http://www.amibroke <http://www.amibroker.com/support.html>
> r.com/support.html> html
> > <http://www.amibroke <http://www.amibroker.com/support.html>
> r.com/support.html>
> >
> > > Yahoo! Groups Links
> >
> >
> >
> > > http://groups. <http://groups.
> <http://groups.yahoo.com/group/amibroker/>
yahoo.com/group/amibroker/>
> yahoo.com/
> > <http://groups. <http://groups.yahoo.com/group/amibroker/>
> yahoo.com/group/amibroker/> group/amibroker/
> > <http://groups. <http://groups.yahoo.com/group/amibroker/>
> yahoo.com/group/amibroker/>
> >
> >
> >
> > > Individual Email | Traditional
> >
> >
> >
> > > http://groups. <http://groups.
> <http://groups.yahoo.com/group/amibroker/join>
> yahoo.com/group/amibroker/join>
> > yahoo.com/ <http://groups.
<http://groups.yahoo.com/group/amibroker/join>
> yahoo.com/group/amibroker/join>
> group/amibroker/
> > <http://groups. <http://groups.yahoo.com/group/amibroker/join>
> yahoo.com/group/amibroker/join> join
> > <http://groups. <http://groups.yahoo.com/group/amibroker/join>
> yahoo.com/group/amibroker/join>
> >
> > > (Yahoo! ID required)
> >
> >
> >
> > > mailto:amibroker- <mailto:amibroker-digest@
> <mailto:amibroker-digest%40yahoogroups.com> yahoogroups.com>
> > digest@xxxxxxxxx <mailto:amibroker-digest@
> <mailto:amibroker-digest%40yahoogroups.com> yahoogroups.com> ps.com
> > <mailto:amibroker-digest@ <mailto:amibroker-digest%
40yahoogroups.com>
> yahoogroups.com>
> >
> > > mailto:amibroker- <mailto:amibroker-
> fullfeatured@ <mailto:fullfeatured%40yahoogroups.com>
yahoogroups.com>
> > fullfeatured@ <mailto:amibroker-fullfeatu
> <mailto:amibroker-fullfeatured%40yahoogroups.com>
red@xxxxxxxxxxxxcom>
> yahoogroups.
> > <mailto:amibroker-fullfeatu
> <mailto:amibroker-fullfeatured%40yahoogroups.com>
red@xxxxxxxxxxxxcom> com
> > <mailto:amibroker-fullfeatu
> <mailto:amibroker-fullfeatured%40yahoogroups.com>
red@xxxxxxxxxxxxcom>
> >
> >
> >
> > > amibroker-unsubscri <mailto:amibroker-
> unsubscribe@ <mailto:unsubscribe%40yahoogroups.com> yahoogroups.com>
> > be@xxxxxxxxxxxx <mailto:amibroker-unsubscri
> <mailto:amibroker-unsubscribe%40yahoogroups.com>
be@xxxxxxxxxxxxcom> com
> > <mailto:amibroker-unsubscri
> <mailto:amibroker-unsubscribe%40yahoogroups.com>
be@xxxxxxxxxxxxcom>
> >
> >
> >
> > > http://docs. <http://docs. <http://docs.yahoo.com/info/terms/>
> yahoo.com/info/terms/> yahoo.com/
> > <http://docs. <http://docs.yahoo.com/info/terms/>
yahoo.com/info/terms/>
> info/terms/
> > <http://docs. <http://docs.yahoo.com/info/terms/>
yahoo.com/info/terms/>
> >
> >
> >
> >
> >
> >
> >
> > _____
> >
> > I am using the free version of SPAMfighter for private users.
> >
> > It has removed 458 spam emails to date.
> >
> > Paying users do not have this message in their emails.
> >
> > Try SPAMfighter <http://www.spamfigh
<http://www.spamfighter.com/len>
> ter.com/len> for free now!
> >
> >
> >
> >
> >
> >
> > _____
> >
> > I am using the free version of SPAMfighter for private users.
> > It has removed 458 spam emails to date.
> > Paying users do not have this message in their emails.
> > Try SPAMfighter <http://www.spamfigh
<http://www.spamfighter.com/len>
> ter.com/len> for free now!
> >
>
>
>
>
> _____
>
> I am using the free version of SPAMfighter for private users.
> It has removed 458 spam emails to date.
> Paying users do not have this message in their emails.
> Try SPAMfighter <http://www.spamfighter.com/len> for free now!
>



I am using the free version of SPAMfighter for private users.
It has removed 458 spam emails to date.
Paying users do not have this message in their emails.
Try SPAMfighter for free now!
__._,_.___

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 NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
http://www.amibroker.com/devlog/

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




Your email settings: Individual Email|Traditional
Change settings via the Web (Yahoo! ID required)
Change settings via email: Switch delivery to Daily Digest | Switch to Fully Featured
Visit Your Group | Yahoo! Groups Terms of Use | Unsubscribe

__,_._,___