Half-cycle differencing for momentum signals

Differencing subtracts a past price from the latest price over a chosen day count. The result is a signed series that can move from one extreme through zero to the opposite extreme. The archive demonstration then asks how that day count should be chosen once a repeating cycle is under study.

A sine-wave stand-in

The working model is a sine wave with period 20 and amplitude 10, oscillating around 100 between 90 and 110. It is used only as a simplified stand-in for cyclic price movement. The dominant cycle is the repeating high-to-high or low-to-low period treated as the reference length against which each differencing interval is chosen and judged.

Leading versus lagging lookbacks

A 5-day difference of that series formed a similar wave at half amplitude that led the original path, so the indicator could turn up while the model series was still falling.

A 10-day interval, equal to half the 20-period cycle, reproduced the original wave with no amplitude loss and with tops and bottoms at the same locations. The demonstration treats one half of the studied cycle as the preferred differencing length because it maximizes indicator amplitude and aligns turns with the modeled market.

A 15-day interval again halved amplitude but lagged the original series, so the model turned before the indicator.

Lookbacks shorter than half the cycle can warn early but lose amplitude and become more prone to random false turns. Lookbacks longer than half the cycle, yet shorter than the full period, lag and can miss much of an extended move. Amplitude loss is that shrinking swing as the interval moves away from half the cycle, which leaves the signal more exposed to random distortion.

The zero-information interval

An interval equal to the full cycle length returns differences of zero and supplies no information about the cycle being followed. That zero-information interval is not a neutral default. It erases the path of the cycle under study.

A momentum rule on the signed series

The momentum rule marks a buy when the differenced series rises after a downtrend and a sell when it declines after an uptrend. Those turns were plotted across the interval variants. The procedure turns the signed differencing series into buy, sell, or stand-aside actions when the indicator itself reverses after an uptrend or a downtrend.

Preset lengths without a cycle check

Applying preset differencing lengths without first checking the data for cycles is described as a hazardous way to use the indicator in trading.

Editorial reading

TradersWeek editorial interpretation, not an archive claim: treat the lookback on a momentum oscillator as a cycle parameter, not a convention. Once a dominant period is identified, the same differencing rule can lead, coincide with, or lag price, and can also collapse to a flat line that supplies no information. The editorial lesson is how to choose and test that interval so entry, exit, and abstention rules stay comparable to an explicit quantitative baseline.