|
PureBytes Links
Trading Reference Links
|
Walter
I organize my worksheets as follows:
1- SP Original
This contains the date, O-H-L-C, volume and OI (only because they come
over automatically using OLE) as well as all of our basic calculations
needed in calculating our indicators.
2- SP TAS
This worksheet uses those calculations generated on SP Original and
calculates our indicator for our original trading system running under
Clipper and TAS. It uses the prices from the first worksheet as well as
some of the more basic calculation.
3- SP MS
This worksheet uses the basic calculations created on worksheet 1 and
modifies our momentum calculations to agree with what we have running in
Metastock. It also uses whatever preliminary calculations required from
worksheet 1 or 2 above, in order to eliminate any redundant calculations.
4- SP Summary
This is where I post the results on a worksheet. It includes date,
O-H-L-C, Contrary (one of our indicators), and the calculated TAS, MS and
Bonds indicators. Bonds come from another worksheet used to calculate our
Bond indicator (same calculations as SP MS but for Bonds). This spreadsheet
also includes the summary results of our intermediate calculations. With
this information, I am able to maintain an ongoing, updated status of 4
different systems. These are:
a) Our original S&P trading system which includes requiring that Bonds at
least not signal a trade in the opposite direction. We've used this system
for the last 12 to 16 years with minor modifications. (62% 376.90 points
profit)
b) The original S&P trading system with the bonds disconnected. I noticed
last year that bonds seem to have decoupled from equities as far as our
systems went, for maybe the second time in 16 years. I posted this as a
question to the group last year to see if anybody else had noticed this. I
started maintaining this system (sans bonds) and it consistently generated
better profits than our original system. (68% 489.55 points profit)
c) A manually posted signal for our intermediate trend, on those dates
that it occurs. Substantially less activity (currently long from 10/1/98)
than a or b above, so it was easier to post manually until I add the
intermediate trend signal to the worksheets. We're constantly experimenting
in this area, so sometimes it's easier to let my dad keep creating new ideas
under Clipper and then manually posting them to this worksheet. (88% and
467.50 points profit)
d) A modified signal requiring both b and c above and trading on the day
they both come into agreement. (100% and 565.10 points profit)
5- Trades
This is a summary worksheet where I post all real trades.
The percentages and points shown above by each 'system' represent all
activity from 5/1/98 to yesterday, and include any profits and/or losses on
open trades through EOD Thursday, February 24. I didn't actually go back
and count all of the trades for a and b since I've automated the
calculations based upon whether a trade is profitable or not. I take the
total number of all positive results and divide that by the total number of
results (positive and negative) to develop the percentage of profitable
trades. The points total is a total of the profits and losses for that
system, and where applicable uses real fill data. I also monitor the
improved profitability of the bondless system versus the system with bonds,
so that I can determine when the marketplace returns to it's more 'normal'
relationship between bonds and equities and when we should return to our
original trading system.
I really need to learn more about VBA in order to move on to the next level
of sophistication. These are the next things I would like to add to Excel.
1- Direct access to the MS data via MS File Library to eliminate the OLE
facility. That capability is currently not working and neither is MS 6.52.
I still use MS for DOS (v4.5) for my MS calculation. I haven't had time to
go play with OLE, so I think I'm going to just move on to a more direct
access.
2- Automating the actual signal. Right now, the spreadsheet is updated
automatically with the results from the above mentioned calculations. I
then manually enter all 4 signals.
3- Incorporate systems testing capabilities. We current do all of our
research within Clipper since my dad is more comfortable there. When we
converted over to microcomputers (in the days of CP/M and home built micros)
from mainframes (around 1972 I believe, but way before there was an Apple or
IBM PC), I taught my dad dBaseII and how to try out different ideas using a
few dBase templates I created for him. He's been doing it for years, all
the way through the transition from dBaseII, dBaseIII and finally Clipper
(an xBase derivative compiler language). What I need to do is create the
same sort of template capability within Windows which will let him test our
new ideas, etc. He's 90 years old (come June) and I'm not sure how he'll
take to learning a new methodology. Might have to keep the old Clipper
stuff running to keep him running. :)
Anyway, that's sort of an overview of what we have running now. My
development progress has been slow lately due to some health problems, but
I'm getting back up to speed on VB and will hopefully get moving. Once I
complete all of the grunt work above, I'll be able to get back to developing
and designing new indicators (like we need more <G>).
What we have found is that the marketplace is fluid and that you need to
constantly monitor what you do to stay on top of it. What worked for the
last couple of years might turn to poop overnight. I think a lot of this is
caused by more computers performing number crunching than ever before.
Years ago, when everything was done by hand or Freiden calcu1ators (way
before the hand calculator even), when you developed a methodology, it
stayed viable for a fairly long time. Then, more and more people started
using hand calculators, etc. and the life expectancy of a signal started to
decline. That's why we concentrated on developing our own indicators. We
recognized that everyone would eventually be able to use all of the more
common ones, and work effectively with them using computers that were
becoming more cost effective every year. Eventually, through trial and
error, they would be able to identify what indicators were working at a
point in time, and then overtrade them to the point that they become
worthless or less valuable over time. We were lucky that my first job after
college was with IBM, so I learned computers early on (my first one in 1960
in college) and was able to have access to main frames to test out our
ideas, etc. even before most investment firms were using computers to
process statements, let alone analyzing market trends, etc.
Anyway, that's why we spent all of the time and effort to develop our own
indicators. Even so, eventually someone comes up with one of them (or
fairly close to it) on their own. Larry William's %R is quite similar to
one of ours that we had been using for 20 years prior to his stumbling on %R
(G). Now, this is just one component of one of our indicators, so the
impact on our system from this being made public was negligible.
Anyway, that enough rambling for now. I've got to get on the treadmill and
get to work around here.
Good luck
Guy
-----Original Message-----
From: owner-metastock@xxxxxxxxxxxxx
[mailto:owner-metastock@xxxxxxxxxxxxx]On Behalf Of Walter Lake
Sent: Friday, February 26, 1999 4:28 AM
To: metastock bulletin board
Subject: Excel organizing
Hi Guy
Thanks for your note.
Any comments on how you personally like to "set up" or organize sheets or
workbooks from a programming point of view. How do you separate "things" and
"actions" to keep them all "straight".
I use: data --> variables --> trading system or indicator tests including
reports
Best regards
Walter
|