|
PureBytes Links
Trading Reference Links
|
Yuki,
This is a very interesting point which I brought up some time ago in
an unrelated thread, which pretty much fell on deaf ears at the time,
discussing same day trading systems and the implications of
datastreams use for testing etc. The fact of course is that what is
reported as the closing price of the day typically does NOT happen
during the market day but after its over i.e. the settling price
which of course we don't get to trade at.
--- In amibroker@xxxxxxxxxxxxxxx, Yuki Taga <yukitaga@xxxx> wrote:
> Hi UM,
>
> Tuesday, July 8, 2003, 11:17:02 AM, you wrote:
>
> umtod> The same question you could ask everybody offering any
> umtod> (financial) products. Why not keep it secret.
>
> The answer is pretty apparent to me: They can make more money
selling
> the system than trading it. (The system looks better than it really
> is.) Sorry, but the mathematics of selling a good system just don't
> work out. Way too much hassle, and even legal liability. Some
> people are not going to follow it, and take a lot of your time to
> complain. Others may actually sue you. If you have a good system,
> raising capital is a piece of cake. But not by selling it. ^_^
>
> umtod> My current sytems work superior if the following framework is
> umtod> given: Enter a position some minutes before market close and
> umtod> exit the position on the next day or days again some minutes
> umtod> before market close.
>
> Huh???
>
> How can you back test on such subjective criterion? You are back
> testing EOD data, as you say below, yet your fill prices cannot
> possibly be determined from any part of EOD data, since they are
> neither open nor closing prices (I won't even mention the absurd
idea
> of filling on a H or L, which I'm sure you're not doing). On that
> basis, EOD data, can you even guarantee that a fill would take place
> "some minutes before market close"? I kind of doubt it, as "some
> minutes before market close" is famous for "some unexpected
> volatility". I have seen stocks, deeply liquid high-caps, with very
> high range bars on the day in question, be at 98 percent of the
day's
> range 10 minutes before close, 50 percent of the day's range 5
> minutes before close, and close on the high (100 percent of the
> range) -- or the low (zero percent of the range, but I will admit
> this is rare, although it happens). You may go in with limit orders,
> but your system, testing on EOD data, would seem to have no way of
> knowing whether the order was filled or not, since the order is
> neither open nor close. In a back test of course the test engine is
> going to *assume* a fill, because the limit price is between H & L.
> But that doesn't mean anything in the real world. And if you are
> going in with intraday market orders, which guarantee fills, you
have
> nothing to back test. Nothing at all, with EOD data.
>
> umtod> That is: since it was tested on EOD data,
>
> Yuki
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get A Free Psychic Reading! Your Online Answer To Life's Important Questions.
http://us.click.yahoo.com/Lj3uPC/Me7FAA/ySSFAA/GHeqlB/TM
---------------------------------------------------------------------~->
Send BUG REPORTS to bugs@xxxxxxxxxxxxx
Send SUGGESTIONS to suggest@xxxxxxxxxxxxx
-----------------------------------------
Post AmiQuote-related messages ONLY to: amiquote@xxxxxxxxxxxxxxx
(Web page: http://groups.yahoo.com/group/amiquote/messages/)
--------------------------------------------
Check group FAQ at: http://groups.yahoo.com/group/amibroker/files/groupfaq.html
Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
|