Selecting The Best Futures Price Series For Computer Testing by Jack Schwager
One problem that traders studying commodity markets face is the fact that individual futures contracts have price characteristics that are not continuous with other contracts within the same market. Jack Schwager, author and director of futures research at Prudential Securities, has some suggestions on dealing with this problem.
System traders who wish to test their ideas on futures prices face a major obstacle: the limited life span
of futures contracts. In contrast to the equities market, where a given stock is represented by a single
price series spanning the entire test period, in futures, each market is represented by a string of expiring
contracts. Proposals for a solution to this problem have been the subject of many articles and much
discussion. In the process, substantial confusion has been generated, as is evident by the use of identical
terms to describe different types of price series. Even worse, a great deal of misinformation has circulated
on this subject. My goal here is to attempt to set the facts straight.
Four basic types of price series can be used. Each has advantages and disadvantages.
ACTUAL CONTRACT SERIES
At a glance, the best apparent route might be to simply use the actual contract series. However, two major
problems arise with this approach. First, if you test a system over a meaningful length of time, each
market simulation requires a large number of individual price series. For example, a 15-year test run for a
typical market would require using approximately 60-90 individual contract price series. Moreover, using the individual contract series requires an algorithm for determining what action to take at the rollover
points. As an example of a problem that may develop, it is quite possible for a given system to be long in
the old contract and short in the new one or vice versa. These problems are hardly insurmountable, but
they make the use of individual contract series somewhat unwieldy.