|
PureBytes Links
Trading Reference Links
|
Guy
I see your methods. I like to use Explorer to search through the entire
market of stocks trying to find which ones to buy and when, when to sell and
move into another stock and when to go long or short. Would your methods
handle this or how would you go about it?
neo
-----Original Message-----
From: owner-metastock@xxxxxxxxxxxxx
[mailto:owner-metastock@xxxxxxxxxxxxx]On Behalf Of Guy Tann
Sent: Friday, October 13, 2000 4:02 PM
To: metastock@xxxxxxxxxxxxx
Subject: RE: New Buy
Dave,
First, I apologize for the delay in getting this out, but somehow it ended
up in my draft folder and never got sent.
Here are some of our thoughts on this.
When developing our own systems, even though we have used computers since
1961-2, we test out all of our indicators by hand. This provides us with
the following:
* A better feel for the indicator. This "feel" translates into the ability
to "fine tune" your rules and try different weights, combinations, etc. and
other techniques that you might have in your tool box like binary waves,
moving averages, etc.
* This will also give you the opportunity to look at the signals being
generated to determine how sensitive your indicator is and whether or not
you'll need to slow it down or try to speed it up.
* A manual review of the indicator going back however many years you are
comfortable with. We have a disagreement between ourselves as to what sort
of back testing we should do. On the one hand, my brother likes to go back
to 1987 or earlier in checking out a new signal. My dad and I, OTOH, feel
that going back 5 years or so is enough, as we are more interested in what
is working now as opposed to what worked 15 years ago. Again, we're still
talking about doing this manually. We use Excel for this. We extract the
database from MS and load it into an Excel spreadsheet. We enter our
signals manually and develop profitability percentages, overall
profitability, average profit per winning trade, average loss per losing
trade, etc. We do this on an annual basis.
Once we have determined that we have an indicator that looks interesting,
we then will automate the testing procedure. The reason we resort to
computers at this stage is to eliminate any positive or negative bias we
might have entered into the equation with wishful thinking, for example.
You can't lie to a computer like you can to yourself, saying I really
wouldn't have made that trade, or vice versa. Like my latest S&P buy
yesterday. Until the bottom fell out today, I was ahead. At least my
stocks held their own today and that's a positive sign. :)
After developing your trading rules, and testing them, we might use Equis'
optimizing features to see if we can improve on our indicator. Now here is
a small warning. If we find too big an improvement through optimization, we
will go back to step one and look at everything in more detail since we're
supposedly talking about fine-tuning here.
Guy
Politicians and diapers have one thing in common. They should both be
changed regularly and for the same reason.
-----Original Message-----
From: owner-metastock@xxxxxxxxxxxxx [mailto:owner-metastock@xxxxxxxxxxxxx]On
Behalf Of Dave Hembree
Sent: Thursday, October 12, 2000 11:59 AM
To: metastock@xxxxxxxxxxxxx
Subject: Re: New Buy
----- Original Message -----
From: "Guy Tann" <grt@xxxxxxxxxxxx>
To: <metastock@xxxxxxxxxxxxx>
Sent: Wednesday, October 11, 2000 3:44 PM
Subject: RE: New Buy
Guy......Could you breifly explain to a newbie working on his own trading
system what you mean by "optimization trap", and developing a system using
100% hindsight"? I hope I haven't fallen into that trap, but advice from
someone such as yourself I believe would clear it up for me.
Thank You, Dave
> I've tried to encourage everyone to work on developing his or her own
> trading system as we believe that anything you can buy is not worth what
you
> pay for it. We also encourage everyone to paper trade for a while before
> risking real money, and finally we recommend that you start this
development
> manually, just so you don't fall into the optimization trap, where you
> develop your system using 100% hindsight.
>
> Regards,
>
> Guy
>
> Politicians and diapers have one thing in common. They should both be
> changed regularly and for the same reason.
>
> -----Original Message-----
> From: owner-metastock@xxxxxxxxxxxxx
[mailto:owner-metastock@xxxxxxxxxxxxx]On
> Behalf Of neo
> Sent: Wednesday, October 11, 2000 12:09 PM
> To: metastock@xxxxxxxxxxxxx
> Subject: RE: New Buy
>
> Guy
>
> Would you mind sharing the formulas and system?
>
> neo
>
> -----Original Message-----
> From: owner-metastock@xxxxxxxxxxxxx
> [mailto:owner-metastock@xxxxxxxxxxxxx]On Behalf Of Guy Tann
> Sent: Wednesday, October 11, 2000 11:58 AM
> To: metastock@xxxxxxxxxxxxx
> Subject: RE: New Buy
>
>
> Leo,
>
> H700 is one of our Intermediate Term Signals (ITS). Our system is
comprised
> of two parts. The first half being a short term signal (which we call
SP39
> which simply stands for the 39th indicator in MS) and the other half being
> comprised of 4 ITS (H700, HSIG, H66R and HGGG and don't ask me how our dad
> comes up with some of these names). :) All of these ITS indicators
> (signals) run in DOS as I haven't been able to convert them into MS yet.
>
> Guy
>
> Politicians and diapers have one thing in common. They should both be
> changed regularly and for the same reason.
>
> -----Original Message-----
> From: owner-metastock@xxxxxxxxxxxxx
[mailto:owner-metastock@xxxxxxxxxxxxx]On
> Behalf Of leo.timmermans.lt@xxxxxxxxxxxxxxxx
> Sent: Wednesday, October 11, 2000 1:40 AM
> To: - *metastock@xxxxxxxxxxxxx
> Subject: Re: New Buy
>
> Hello,
>
> What is this H700 ???
>
> Kind regards
> Leo
>
>
|