|
PureBytes Links
Trading Reference Links
|
Ditto that: I agree that backtesting done with Tradestation should be taken
with a grain of salt. I don't trust it. I seen too many times when it
should have done something but didn't or visa versa. I prefer judging for
myself, the validity of a signal and system. That's not to say EL is not
useful.
I find it useful for automation.
-----Original Message-----
From: Dennis Holverstott [SMTP:dennis@xxxxxxxxxx]
Sent: Thursday, March 05, 1998 11:38 AM
To: omega-list@xxxxxxxxxx
Subject: Re: Problem Exiting at End of Session??
> > Realtime, you would know the session was ending and exit.
>
> How would you do that in T/S, in realtime, on the days that
> end before SESS1ENDTIME?
I guess, if the built-in exit MOC stop wouldn't do the job (should in
most cases), you could always add an input for the session end time and
use that rather than the built-in sess1endtime. It would work realtime
cuz you could change the value of the input. Wouldn't work for
backtesting though.
IMHO there are so many holes in the way TS handles orders that systems
are only good for backtesting and then the results should be taken with
a grain of salt. For realtime trading, you're better off coding them as
indicators with alerts and print logs and using your judgement about
whether a signal is "real".
A number of people, including Bob Brickey and Mark Brown, have posted
real-world scenarios where TS either gave a signal it shouldn't have or
didn't give a signal it should have and, when they reloaded the chart,
the signals were different. If you make liberal use of the print log and
evaluate what it says before jumping into a trade you *might* be able to
catch some of these errors. No guarantees but it's better than just
blindly following the little arrows on the screen.
And don't even get me started on the problems involved with trading
multiple contracts. :-)
--
Dennis
|