|
PureBytes Links
Trading Reference Links
|
----- Original Message -----
From: "bex1210" <hmw3@xxxxxxxxxxxxxx>
To: <equismetastock@xxxxxxxxxxxxxxx>
Sent: Monday, October 06, 2003 12:46 PM
Subject: Re: [EquisMetaStock Group] MaxLocksperFile
> Tom:
>
> Thanks for the feedback. A few more questions:
>
> 1. Did you have any problems with other programs by raising the MaxLocks
to 500,000?
No, could never find any other problems. And during the day I run 2
instances of eSignal, 3 IE explorer browser windows, my broker's window,
Outlook express, and Excel all at once. Sometimes also have Metastock and
or Powerpoint or Word open as well. Have 1GB ram.
>How large of a ST_Data.mdb file can you now create before you get the error
message?
Don't know. Once I stopped saving the results, started running quick test,
etc etc, my problems became minimal and livable.
> 2. I never thought of exporting to Excel, and storing it there. Exactly
what gets exported, >and how does the exported file size compare to the size
of the ST_Data file? Do any >graphics get exported, or just the data?
I just save the results in the Summary Window that pops up first. For
example if I'm running an optimization, I highlight all the lines, right
click on copy and then go to Excel and paste. I don't save graphics or
trades or anything else. Often I'll also highlight one of the results and
click "view". Then I'll click "System" and copy the buy/sell rules and
paste into Excel right along side the results. Then in Excel I will make
notes to myself as to my observations on the test.
When developing a system I break the total time period into pieces. For
example I just finished one that used 4 1/2 years of data. Actually the
last 6 months were not used for optimization. They were only used as a walk
forward test at the end. The four years were were broken into 4 time
periods and the optimizations were run on each of the periods. They were
copied into Excel in a horizontal fashion left to right. Once in Excel I
could sort each time period for parameters however I wanted. By doing this
you can see if there are any parameters which worked in all time periods.
Also in Excel you can do things like 3D surface analysis and plot your own
stuff.
You should experiment a little bit and save whatever makes sense.
No graphics get exported because I don't think there are any in a quick
test, nor to I care.
When something looks reasonable I will go back and run full system test,
look at equity curve, plot signals, etc etc.
The system I developed has several several hundered runs and I believe the
Excel file got to be about 15MB.
I used to save lots of stuff but found I never used it anyway. I found
that it was much more valuable to save a little, and even more important to
save a note along with it as to my impressions of the test. How often do
you go back and look at a test? When you do a couple weeks later do you
even know what the heck the test was or what it meant? Too bad MS doesn't
allow you to add a note to the system test on conclusion. That would be a
nice touch.
> 3. I now run all of my tests in "Quick Test" mode, but I still have the
errors. The Metastock technician told me that the size of the data file was
the same; only the report file was smaller with "Quick Test". Have you
confirmed that there is a real size difference in ST_Data?
No I haven't confirmed the size is any different. Just asssumed it was
smaller because you can't see as much data. But most of my problems went
away with combo of quick test and regular deleting of tests after 5-10 at
most.
>
> Harry
>
> --- In equismetastock@xxxxxxxxxxxxxxx, "Tom Sprunger" <tlsprunger@xxxx>
wrote:
> > I gradually raised mine. It is now at 500,000 (WinXP) and I still get
> > occasionally the problem.
> >
> > I have had the best results (fewest problems) by being dilligent in
> > discarding runs or exporting the results to EXCEL. I realized that most
of
> > the runs were not worth keeping and just a part of the development
process.
> > So I export the results immediately to Excel, along with the description
of
> > the system test/parameters were. Then record my thoughts in Excel, like
> > "didn't work, looks like the trigger was too high, try using higher
> > trigger". I found it helpful because you always tend to make some sort
of
> > judgement immediately upon examining the results. Plus once results
are in
> > Excel you can perform statistics on the results that you can not do in
> > Metastock. If you have recorded this, you can always rerun the system
test.
> >
> > I also run the Clean utility every few runs.
> >
> >
> >
> > ----- Original Message -----
> > From: "bex1210" <hmw3@xxxx>
> > To: <equismetastock@xxxxxxxxxxxxxxx>
> > Sent: Sunday, October 05, 2003 11:52 PM
> > Subject: [EquisMetaStock Group] MaxLocksperFile
> >
> >
> > > I have Mdtastock 8.01 EOD, running on Windows 98SE. I have recently
been
> > running larger System Tests, and now when I go to Discard a test get a
> > Program Error with the message: "file sharing lock count exceeded.
Increase
> > MaxLocksperFile."
> > >
> > > I contacted Metastock support, and they showed me how to change this
> > setting in the Registry. The "default" is 9500 locks. It was suggested
to
> > double or triple this. The technician I spoke to has his set to 250,000,
but
> > he also has Windows 2000.
> > >
> > > I did some checking, and apparently a very high setting can slow down
> > execution of some programs. I first increased mine to 20,000, and then
to
> > 40,000. If I run a System test with 100 stocks, 4 years of daily data,
and
> > 15 optimization settings (1440 tests in all), I still get this error
message
> > when I Discard the file. If I run it with 8 optimization settings, I can
> > discard OK. (I select Discard and Clean-up after every system test, so
their
> > are no residual tests which Metastock is also trying to discard.)
> > >
> > > The data for the system test is stored in ST_DATA.mdb. A file that is
43
> > Meg Discards OK. A file that is 71 Meg creates the error message.
> > >
> > > Has anyone experienced these same problems, and if so, what
> > MaxLocksperFile settings have been successful?
> > >
> > > Having to run optimization tests with such a small number of variables
so
> > as not to trigger this error message is really labor intensive.
> > >
> > > Harry Ward
> > >
> > >
> > >
> > > To unsubscribe from this group, send an email to:
> > > equismetastock-unsubscribe@xxxxxxxxxxxxxxx
> > >
> > >
> > >
> > > Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/
> > >
> > >
> > >
>
>
>
> To unsubscribe from this group, send an email to:
> equismetastock-unsubscribe@xxxxxxxxxxxxxxx
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/BefplB/TM
---------------------------------------------------------------------~->
To unsubscribe from this group, send an email to:
equismetastock-unsubscribe@xxxxxxxxxxxxxxx
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
|