|
PureBytes Links
Trading Reference Links
|
I've played with what are essentially spread trade ideas for some time and
as there is some current discussion and interest on the list about spreads I
thought I'd share a few thoughts. I consider the paradigm of what I've done
as forcing opportunity. That is, trying to manufacture (very) attractive
risk:reward trades. Here's a conceptual example.
When hedging a portfolio there is always a tracking error, which in and of
itself can be considered a trading opportunity as it oscillates within very
predicable channels. If it doesn't then the hedge should be better
constructed. However, constructing a hedge with an intentional margin of
error obviously causes a greater tracking error, thus widening the channel
(of oscillation) and therefore creating a more tradable opportunity.
Using the Dow 30 index to hedge a portfolio of the Dow 20 (any 20 stocks of
the index), a hedge may be constructed using the relative beta of the
portfolio to the index to arrive at a hedge ratio. Dividing the volatility
of the index by the volatility of the portfolio to arrive at a relative beta
is preferable than using price, for me at least. Oscillation of a hedge
ratio may be thought of as the representation of a hedge tracking error. A
strategy that trades derivatives of the index and/or derivatives of the
portfolio where trades are signaled as the tracking error reaches over or
under the boundaries of the oscillation-channel seem to have merit. However,
if the hedge is well constructed, the opportunity will likely be quite
small.
The objective of forcing an opportunity may be thought of in terms of
increasing the oscillation of a tracking error. There are several choices
one can make to in effect miscalculate a hedge ratio so as to arrive at an
acceptable risk:reward ratio. The focus I've used is to increase reward more
than reduce risk as the latter is inherently reduced through hedged
diversification. An obvious approach is to design or use an optimized system
as the parameters can be almost totally quantified.
To say the least, I've found TradeStation quite inadequate for this kind of
work. Still, an approach that I've used is to send signals from one system
to other systems using (the old) globalvar.dll that Doug Demming wrote some
years ago, and then consolidating the second level systems to arrive at
performance results. It's very cumbersome and can't extend to all of the
derivatives from which one might choose to implement these kinds of trades.
Another issue for which there doesn't seem to be an "intelligent" tool is
monitoring positions and risk. If a spread trade is to include stocks,
futures and options and there are several positions, using a spreadsheet is
contingently fraught with error.
In my travels if you will, I've actually discovered "arb shops" that do what
I've outlined, but with considerably more complexity on an intraday basis.
I've also observed more than a few traders managing risk and keeping score
with a pencil and paper! An "old guy" who founded a BD that now employs 2K+
people said, "If I can't do it on the back of a matchbox, it's too
complicated." I think the matchbox is only a point of focus for amazing
mental calculations. The point is, I'd suggest that spread trading where the
opportunity is forced is a lucrative strategy. But with better tools than
what TS offers, one could really go to town.
Colin West
|