|
PureBytes Links
Trading Reference Links
|
--- In amibroker@xxxxxxxxxxxxxxx, "Fred" <fctonetti@xxxx> wrote:
> Let me ask another question then ... Assuming for a minute that the
> system trades both long and short would you allow some candidate to
> qualify to be traded long OR short only based on the fact that it
had
> done well across the different market environments ( bull/bear/flat
)
> on the long side but not on the short side or vice versa.
Absolutely! I'd take it a step further and say the system could start
out as long only to begin with. As long as it trades well enough to
meet the criteria. This goes against some paradigms but paradigms
have wasted years of my time.
> With regards to the area that you find of interest, I don't think
I'd
> qualify this as optimization of robustness. However for example if
> in the initial steps we had picked some relatively recent time
> periods to use as bull/bear/flat market segments to test over and
> then select issues based on their performance during those periods
> using whatever yardstick one happened to like, then might it not
make
> sense as time rolled forward to replace an older bull, bear or flat
> segment of market that had been used to test system robustness and
> perform issue selection with a newer segment of the same type to go
> through at least the issue selection process again ?
When you say issue selection, are you talking about the nomination or
the confirmation part? If confirmation, it would be better to *add*
the new data in as they become available because criteria 3-5 (when
used for robustness or confirmation) inherently work better (within
limits) with more data. If nomination (for example using an
algorithm)maybe but I doubt it because I've never been able to wring
much advantage out of OOS techniques (with data that exists) and that
sounds like a twist on OOS if I understand you correctly. But still
worth looking into. I've been surprised before.
>
> --- In amibroker@xxxxxxxxxxxxxxx, "quanttrader714"
> <quanttrader714@xxxx> wrote:
> > --- In amibroker@xxxxxxxxxxxxxxx, "Fred" <fctonetti@xxxx> wrote:
> > > The implication here then is that the best behaved issues of
> yester-
> > > year will continue to be so ad inifitum.
> >
> > More precisely, it's the consistency of the *interaction* of the
> > robust system with the best behaved issues that we're counting on
to
> > continue into the future.
> >
> > >Given that any stocks chart
> > > of the past could virtually look and trade like any other stock
> in
> > > the future one would think that some more robust ongoing
> > methodology
> > > of issue selection would be more likely to succeed going
forward.
> > I
> > > know there are some pretty large generalizations here but this
it
> > >how it "feels".
> >
> > This is an area of great interest to me, one that I've looked
into
> on
> > and off in the past but have recently dusted off again. Not an
easy
> > nut to crack because if you buy the interaction premise, the
> > methodology could conceivably be different for each system (let's
> > optimize robustness, lol).
> >
> > >
> > > --- In amibroker@xxxxxxxxxxxxxxx, "quanttrader714"
> > > <quanttrader714@xxxx> wrote:
> > > > No, I don't. The idea is first to run all 5 criteria to see
if a
> > > > system is robust(by this definition anyway). Once "deemed"
> > robust
> > > you
> > > > then try to trade it on the best behaved issues. Seems to me
> that
> > > > this makes sense no matter how robust the system is, because
> > > *nothing*
> > > > short of what God could create will trade everything well
IMO.
> > You
> > > > don't even have to use the S&P stocks to find the best
behaved
> > > issues
> > > > if you're overly concerned about curve fitting, although I
> think
> > > this
> > > > would unnecessarily eliminate tons of great candidates.
> > > >
> > > > As for the process of finding the best issues, even if you
use
> an
> > > > algorithm I personally would still want to manually run them
> > through
> > > > criteria 3-5 as a double check. So 3-5 have utility beyond
> simply
> > > > being part of a robustness standard, at least the way I use
> them.
> > > >
> > > > I think everyone knows this but I like to keep repeating it:
> this
> > is
> > > > only one way to skin the cat, certainly not the only way. I'd
> > add
> > > > that it's also a very labor intensive and boring way too
because
> > > > there's a lot of manual checking and it looks mostly to hit
> > singles.
> > > >
> > > > --- In amibroker@xxxxxxxxxxxxxxx, "Fred" <fctonetti@xxxx>
wrote:
> > > > > Ergo my question about dd's. I didn't think it needed much
> in
> > > the
> > > > > way of interpretation except to understand what context the
> DD's
> > > > were
> > > > > in.
> > > > >
> > > > > As a followup here I have a question. If the next step(s)
are
> > > > issue
> > > > > nomination/selection and we chose to do that manually as
> > opposed
> > > to
> > > > > having some algorithm for objective selection do you feel
> that
> > > then
> > > > > has the potential for calling into question the robustness
of
> > the
> > > > > system which now incorporates issue nomination/selection ?
and
> > if
> > > > not
> > > > > why not ?
> > > > >
> > > > > --- In amibroker@xxxxxxxxxxxxxxx, "quanttrader714"
> > > > > <quanttrader714@xxxx> wrote:
> > > > > > Nobody's taken a stab at interpreting the pics I posted in
> > the
> > > > > example
> > > > > > folder of the photo section, so here goes... This is a
good
> > > > system
> > > > > > paired with a bad issue... it's my benchmark system
> > > (w/commission
> > > > > > setting of 1%) which is robust but it certainly doesn't
> trade
> > > > > > everything well, lol. Proof positive from this stock!
> > Looking
> > > at
> > > > > pic
> > > > > > 1 (% profit/trade), winners look fairly consistent but
> losers
> > > are
> > > > > not
> > > > > > and are nasty on top of it. Pic 2 gives the opposite
view
> > > (losers
> > > > > > look more consistent than winners) which is why criterion
3
> > > > requires
> > > > > > that both be consistent. In this case, the likely
culprit
> of
> > > the
> > > > > > disagreement (you'd have to actually dig into the trades
to
> be
> > > > sure)
> > > > > > is the system couldn't get out of the losing trades this
> stock
> > > > threw
> > > > > > at it quickly enough. Even so, the probability of profit
> > after
> > > 10
> > > > > > trades is not bad (pics 3 & 4) although I'd feel much
more
> > > > > comfortable
> > > > > > with 90%-95% as opposed to 80% (any questions on reading
> > > these?).
> > > >
> > > > > The
> > > > > > potential drawdowns are *awful* (pics 5 & 6). For
example,
> > 80%
> > > > > > likelihood of drawdowns 30% or greater, 50% likelihood of
> > > > drawdowns
> > > > > > 40+% or greater, and 10% likelihood of drawdowns 70% or
> > > greater.
> > > > > > Egad!!!!!!
------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/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/
|