|
PureBytes Links
Trading Reference Links
|
Now that we've look at an example of timing and selection, the
logical question is whether they can be applied to stocks, ETF's, etc.
1. Timing - if you have a useful small-cap fund signal, it can
obviously be applied to such stocks. Several of us have PT and PB
models that incorporate it.
Chuck had asked there is money to be made in trading ETF's. I think
that there is in some cases. Let me give a simple example that I
use. During the "buy" times of a signal, the Russell growth
outperforms the Russell Value. As you probably know, the Russell
2000 can be segmented into these two components. Two ETF's, IWO and
IWN track these components. They have marginally adequate volume and
spreads. Actually, there is something more interesting that can be
done if your brokerage can short the IWO during "sell" periods in a
spread with long IWN. Value tends to outperform growth during the
sell periods and the spread is low risk.
2. Selection - as Fred has noted, NCALPHA is a useful strategy for
ranking some classes of stocks. I know that Gary has been doing some
work in this area also. This is more of an intermediate strategy as
Gary explained in his screen cam movie.
Obviously, the current events issue is the various types of trading
limitations that are placed on mutual funds. Many of the higher
Alpha small-caps have either frequent trading restrictions or exit
fees during the first 3-6 months. One example of the extreme cases
is the BRSIX fund. If you sell within a short time of your buy, you
will be banned from buying in the future. The way that most of the
FT/Trade group is dealing with this is with hedging.
Hedging is done by determining a hedge ratio that neutralizes the
market component of the fund during the sell times. For you MPT
types, this leaves you to make a little money on the positive Alpha
of a good fund. This is done by first determining which Rydex or
Profunds or Potomac fund is most highly correlated with the target.
Then the hedge ratio is determined by maximizing the UPI of a trading
result based on a signal. An augmented hedge is used to do this.
Here's the idea. Assume that you have a small-cap fund X and the
hedge ratio is 60/40. During the buys, you hold 60% X and 40% of a
long Russell funds (UAPIX which is 2x, if you have a really good
signal). During the sell times, you switch the 40% to a short
RUssell funds such as POSSX. Because Rydex, Profunds, and Potomac
encourage short term trading, you can do this very frequently. The
only issue is that some brokerages impose a 1 day settlement, so it
is better to do this with money at Rydex, Profunds, Potomac.
Hope all of this helps-
Bruce R.
--- In amibroker@xxxxxxxxxxxxxxx, "bruce1r" <brucer@xxxx> wrote:
> Let's get the question of system performance out of the way first.
> Most of the more useful FT signals have been developed in Trade.
> Most have been backtested into the mid 90's (although some go back
to
> 9/1/88 - the start of the FT database). The signals tend to fall
> into two categories - basic and aggressive. Please understand that
> I'm offering a a couple of data points only for illustration.
>
> 1. Basic timing -
>
> The basic metric is a signal called RUTVOL. It is a Russell 2000
> based timing signal with volume confirmation. It has not been re-
> optimized for many years. Here are the results from 9/1/88 -
11/5/03
> with 1 day trading delay for the Russell to money market -
>
> Russell 2000 held - 9.05% CAR
> Russell 2000 / RUTVOL - 16.63%
> HRTVX (Heartland Value) / RUTVOL - 21.83%
>
> Many of the aggressive small-cap's didn't start until 94-97. One
> example is Fremont Micro-Cap. Here are the results from 7/6/1994
to
> 11/5/03 with 1 day trading delay to money market -
>
> FUSMX (Fremont MicroCap) / RUTVOL - 30.02% MDD=19%
>
> 2. Aggressive timing -
>
> Here we are getting into the question of acceptable optimization,
> robustness, etc. But one of the more robust, aggressive signals is
> called FUBAR5. Some might find is premise and construction
> controversial, but that is a story for another day. The results
are -
>
> FUSMX / FUBAR5 - 41.41% MDD=8.7%
>
> A variation of it that uses shorting during longer term bear
trends,
> achieve significantly more.
>
> Another major toolset deals with selection. Admittedly, I'm about
to
> gloss over some issues such as survivorship, and others, but let's
> say that we used NCALPHA for fund selecton instead of just trading
> FUSMX in the example above. The parameters are - hold 3 funds, min
> hold 5 days, max rank held = 6.
>
> FUSMX / RUTVOL - 34.8% (compared to 30.02 w/o selection)
>
> Now, can you use these signals for other things, such as stocks,
> ETF's, etc. Yes. More in the next post.
>
>
> --- In amibroker@xxxxxxxxxxxxxxx, "bruce1r" <brucer@xxxx> wrote:
> > I thought that I offer a primer on FT then get to some specific
> > trading questions. First, let's separate the FastTrack world
into
> > four topics -
> >
> > 1. Mutual fund trading
> > 2. FT as a tool
> > 3. Trade
> > 4. Community
> >
> > Mutual fund trading may seem conservative to some, but it is just
> > another class of issues that range from conservative to
> aggressive.
> > Much of FT evolved out of agressively trading Fidelity Selects.
To
> > that end, the FT program has alway concentrated on pair
switching -
> > trading from one fund to another based on a buy/sell signal.
Long
> to
> > MM is just a special case.
> >
> > FT for Windows is the current incarnation of the tool that is
> bundled
> > with the mutual fund data. FT provides "clean" data on funds
> (which
> > is no small feat) that is back adjusted for distrbutions in a
> timely
> > manner. The tool is a Windows program in architecture, but not
in
> > look and feel. Be that as it may, the tool offers some very
unique
> > capabilities for ranking, manipulation of user defined families,
> > switching, etc.
> >
> > Trade is a language developed Ed Gilbert that is an impressive
> piece
> > of work that is free. It is vector programming language (like
> AB),
> > but with calculations extended to families of funds/stocks, is
> better
> > described as a matrix processor. It is very fast, and allows
> > indicators, timing signals, and trading models to be developed
very
> > quickly. Frankly, until AB, I hadn't found anything close, and
> like
> > Fred, I've used many of them.
> >
> > All of this came together into a community of people from varied
> > backgrounds. They concentrated on market timing and fund
selection
> > and developed techniques that may have lacked the rigor of
> academia,
> > but JUST WORKED ! In 1997, Don Beasley catalyzed this group by
> > developing a "Dominant Market Theory" that focused on the Russell
> > 2000. This turns out to be a very useful tool for timing.
Others
> > expanded on this idea. Don Bell, who is very prolific, has
> developed
> > a myriad of timing signals for small-cap funds, junk bonds,
tech's,
> > etc. Werner Gansz developed the basic Russell signals and also a
> > program called NCALPHA that is the tool of choice for ranking
funds
> > relative to an index based on non-correlated Alpha (which can be
> > thought of as the stock picking skills of the fund manager).
> >
> > Next - how is it used, what are the results?
> >
> >
> > --- In amibroker@xxxxxxxxxxxxxxx, "bruce1r" <brucer@xxxx> wrote:
> > > Chuck - I think that I might be one of the people that Fred was
> > > referring to in an earlier post who comes from this FT/Trade
> world
> > > and also has some experience with Amibroker. I am "stuck"
> > somewhere
> > > for the next day and will have some time to post an overview
and
> > > hopefully answer a couple of your questions. If you bear with
> me,
> > > I'll go until the board tells me to stop. Here are the
bullets -
> > >
> > > 1. The FastTrack community drew a group of people together over
> > many
> > > years who have done a body of very impressive work on selection
> and
> > > intermediate term timing. I've become familiar with the depth
> and
> > > excellent quality of your work, and think that you would find
it
> at
> > > least interesting.
> > > 2. The work was focused on funds, particularly small cap's
> because
> > of
> > > their trend persistence, Alpha, and lower volatility.
> > > 3. Much of the work is transferrable to stocks as Gary S. has
> found.
> > > 4. There are approaches to working around the fund trading
issues
> > as
> > > Fred has mentioned.
> > > 5. For whatever its worth, I've imported/ported much of the
work
> to
> > > Amibroker.
> > >
> > > I'm off to dinner. More later.
> > >
> > > Bruce R.
> > >
------------------------ 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/
|