Are There Persistent Cycles? by E. Michael Poulos
Most technical indicators use a fixed lookback length based on the idea that cycles are present in the price data. Should the same fixed lookback length be used for all markets? Are there persistent cycles present to justify the use of the same lookback length for evaluating the market? Mike Poulos searches through a number of markets for persistent cycles and offers his solutions.
Most technical indicators either smooth (filter) or normalize (on a scale to 0 to 100) the tradeable
price in an attempt to indicate the trend direction or a trend change. The vast majority of technical
indicators use fixed lookback lengths (for example, 14 days for the relative strength index [RSI], 14 and
28 days for moving averages and so forth) in the formulas in constructing an indicator. The popular
justification for these fixed lengths is the supposed existence of persistent cycles. For example, there is
supposedly a 28-day cycle that justifies the use of 14 days (a half cycle) in the RSI oscillator when applied
to all commodity markets. Common sense should alert you to question the existence of any persistent, or
continuously present, cycle.
A problem arises regarding the use of indicators that are using the same lookback length over all time
periods and all markets; these indicators are subject to providing misleading information because the
indicators can be out of step with the markets. Is there evidence of persistent cycles? To answer, first let
us look at close to 1,000 days of closing price data in eight different futures markets and two stocks for
some evidence of persistent cycles by using correlation analysis.