|
PureBytes Links
Trading Reference Links
|
Robin:
>Alex, it's probably true that CSI is the only source of this
>because who else would want to use it? While I'm a big fan of CSI,
>I see zero use for the perpetual contract. I back tested gold one
>time with it and it always stayed long. That's because gold was/is
>a carrying charge mkt and as you move forward in time more weight
>is given to the forward contract and less to the spot month. This
>of course gives the price series an upward bogus bias.
Right, and if you use back-adjusted continuous contracts your bias is
the other way, and the bias is even stronger for very long term
strategies (using 1 year or more time horizons). This is where the
perpetual contract works best.
Perpetual contracts are more useful than any other kind when:
1. Your signals rely on ratios between prices. The usual back
adjustment method can result in negative prices, which make ratios
meaningless. In fact, price ratios become meaningless prior to
the nearest rollover date. Because many securities (including
commodities) have more sensible price changes in percentage units
than price units, having a history that allows you to calculate
percent changes is useful -- especially for portfolios of stocks
where you're more interested in percentage returns than price
changes.
2. You need spot prices to generate signals, and your data vendor
doesn't offer it. Perpetual contracts make an excellent proxy for
spot prices, because the perpetual contract price will never drift
far from the actual spot price. This cannot be said for any other
type of continuous contract.
--
,|___ Alex Matulich -- alex@xxxxxxxxxxxxxx
// +__> Director of Research and Development
// \
// __) Unicorn Research Corporation -- http://unicorn.us.com
|