Percent-offset simple moving average bands with relative-strength confirmation

Two construction choices

The workflow places a moving-average envelope and a relative-strength-gate side by side. Each is a separate construction choice. The envelope answers how a single average of closes is turned into a pair of series. The gate answers whether a relative strength index of those same closes must also hold.

Editorial reading: a later comparison that changes the offset constants, the lookback, and the relative strength condition in one step is no longer testing the same model. Each knob has to stay explicit.

The percent-offset-envelope

The envelope is built from a 15-period simple moving average of closing prices. The upper series is that 15-period average multiplied by 1.02, a 2 percent upward offset. The lower series is that 15-period average multiplied by 0.98, a 2 percent downward offset.

That pair is a percent-offset-envelope: the same lookback is kept, and the average is moved in price space by constants. The number of closes in the mean does not change when the offsets are applied.

Lookback-substitution is a different average

An 18-period simple average is the mean of the last 18 closes and is not a 15-period average shifted upward. A 12-period simple average is the mean of the last 12 closes and is not a 15-period average shifted downward.

The 18- and 12-period pair was proposed by treating 2 percent of 15 as three extra or fewer bars. That step is lookback-substitution. It changes how many observations enter the mean instead of shifting the original 15-period mean.

Editorial reading: the arithmetic that turns 2 percent of 15 into three bars explains why the substituted pair was suggested. It does not make the 18-period mean equal to the 15-period mean multiplied by 1.02, and it does not make the 12-period mean equal to the 15-period mean multiplied by 0.98.

A relative-strength-gate beside the bands

A 14-period relative strength index of closes is added as a separate confirmation series beside the bands. That is a relative-strength-gate: the condition must hold in addition to the envelope comparison, and it is not folded into the average that produces the offsets.

Scan filters and the prior-bar-reference

A long-only scan requires the close above the upper offset series, the prior close below a paired offset series, and the 14-period relative strength index below 40. The lagged close is a prior-bar-reference. It is used so the scan can require a cross rather than only a level.

A short-side scan keeps the same 15-period average and 2 percent offsets but swaps which offset series and prior-bar references enter the filter. The lookback and the size of the percent-offset-envelope stay fixed. Only the side of the comparison and the lagged references change.