|
PureBytes Links
Trading Reference Links
|
...and I thought I'd introduce myself...
I've been using TS4 for about 3 months and joined this list just in time for
the missives of "The Code." (Was the note about a Veeger family in Florida
serious or has no one seen that Star Trek movie?--"It's back and it's gonna
eat us." <G> Clever Code. Go get 'em!)
I was/am a Metastock 6.5 user. MS6.5 provides more built-in and example
code indicators than TS does. But, of course, it's not real-time and there
are other ways TS is clearly a more powerful platform. That's why I'm here.
I'm a contract software designer/programmer by profession. Currently have a
gig that keeps me at home most of the time and this is great because I can
watch the markets.
Did some Globex trading. Read on this list that someone slept near the
computer to hear the alerts: I bought WinBeep and used my wife's work pager
for signals. Slept with the pager in my hand with the buzzer on...Worked
great! Got a page within 1-2 minutes of my alert being hit. This is plenty
of time usually for Globex. Not that I had the smarts or the ability to
trade the signal profitably mind you....Just yesterday learned that my
wife's secretary is contesting all the charges for the text pages I got. My
wife told her that, yes, it must be a mistake, "who would page me so many
times at night." <G>
I had terrific success my first week trading the S&P against the trend (but
I was trying to trade with the trend, go figure). Then lost most of my
gain.
Used Leo-Web. Always filled within a minute bar...Usually get the fill
report within a couple minutes. Incredible. Is there anything better out
there? Hard to imagine.
Decided I'd just watch for awhile. That's been helpful.
I agreed with one of Gwenn's posts that I understood to say something like
the following:
- I'm not smart enough to build a system that should be used without
question over discretion...
- I'm not smart enough to trade without a mechanized alert system that
pulls the indicators I'm interested in together in a coherent way... ==>>
And my belief is (perhaps naively) that alot can be done to build a pretty
good alert system.
Let me digress **********************************************
You have to promise not to laugh....Just bought my first UNRETURNABLE
"indicator software" from a system vendor. The account rep reassured me at
least three times that I'd have access to the code. It cost more than $500
and less than a grand. I had built something myself that seemed to
highlight/count the bars just like the book said it should and that I
thought was close to what this vendor was advertising but I thought "how
could THIS particular vendor charge that much money for something that was
so simple" to code myself. Surely, they must have something that I should
have too. I also bought it figuring I could use the UN-passworded code as a
template for other stuff I wanted to do. It turns out that all the FUNCTION
code is passworded after all. I've been talking with the vendor to get them
to fix this for me. I'll post all the gory details to this list (because I
KNOW it will be of interest) if I don't get satisfaction in one of the three
alternative satisfying approaches I've outlined for the vendor. VENDOR:
Are you listening?
Hey, CODE!!! I figure that if I paid what I did for the logic this
indicator provides that I can't even see and is embarrassingly marginal
(remember, you promised not to laugh at me) I must owe you thousands and
thousands of dollars. Do you take VISA?
End of digression **********************************************
So, before I go on to try to lose more, I'm trying to develop a trading
system based on pretty standard indicators from multiple intra-day
time-frames (Elder's "Factor of Five"--though I like using more time frames
than he discusses--in other words, I think there are multiple wave and tide
periods) that will provide alerts when the indicators line up. Then I'd use
my discretion to trade it.
I read here that someone thought the short list of favorite indicators
couldn't be traded successfully. When I look at my TS screen, I see
something different. But I'm looking at these applied to multiple time
frames and perhaps its this multi-period view that makes them more coherent.
Or perhaps I'm just dead wrong and have a great deal more disappointment
ahead of me...<g>
A set of tests that I want to integrate into the system would provide an
estimate of the "state" of the trends for specific multiple time-frames. I
think I've got the "state diagnosis" functionality conceptually put together
in a way that'll work and the code design done (even if specific "state"
definitions change) for the OB/OS oscillators that have fixed scales. I'm
still refining and debugging the code and wrestling with the 64K ceiling
issue. I only have the beginnings of a concept and code design that
integrates the "state" values for the various time-frames into a
"multi-period indicator state analysis and alert decision" function. This
will take awhile 'cause I'm doing it part-time.
I'd initiate or participate actively in threads that concern:
- how to calculate "state" for OB/OS oscillators without fixed scales. I
have a few thoughts but couldn't initiate this discussion...
- "trend states". (I mean more than below and above and a cross of the
line here.) I'd start this discussion off with some preliminary and I'm
sure crude ideas if two or more pros promise a critique and their insight...
- the concept and/or code design for a "multi-period indicator state
analysis and alert decision" function. Again, I've got crude thoughts but
I'll make time to refine and share them if promised significant/quality
critiques and analysis from pros...
Are these kinds of topics ever discussed on this list? Maybe I'm in the
wrong place. There's ten times more bitching and fighting here than on the
Metastock list....<g>
Steven Buss
sbuss@xxxxxxxxxxx
Walnut Creek, CA, USA
|