Reader tests for unfinished lookback oscillators

A printed reading is still a draft

The relative strength index is a bounded oscillator built from ordered closes over a stated sampling interval and lookback. A printed reading remains a draft forecast until session hours, a plain volatility baseline, and an out-of-sample check are specified.

The commodity channel index is a scaled deviation of price from a moving typical-price mean over a stated lookback. The same unfinished-formula tests that apply to any published study apply here before the output is used as a forecast.

As an editorial matter, those two constructions should not be treated as forecasts until a reader can run the three tests modeled in this correspondence.

Does the sample assume a session the market does not have

The archive records a correspondent arguing that a published gap-based randomness index omitted around-the-clock trading. Continuous markets would look less gappy under that construction, and therefore less random.

Session-continuity is the issue. Markets that trade around the clock can produce nearly gapless series, so a construction that treats gaps as a randomness signal will systematically rate those series as calmer than session-limited series.

The editors accepted the objection and described the index as an unfinished starting point that should be revisited.

Does a noisy reading equal risk

The same correspondent asked whether erratic readings had been tied to risk, and how the index differed from ordinary volatility statistics.

A volatility-baseline answers the second half of that question. A published index should be set beside a plain volatility measure so a reader can see what, if anything, the extra algebra adds.

The archive does not record that those questions were closed. The editors treated the work as unfinished.

Is the construction even documented

A platform user could not obtain an explanation of a bundled support-resistance study from the software vendor and had to hunt for the original construction notes. A support-resistance study is not usable as a model until its construction and intended reading are documented independently of the platform label.

A data user reported that many vendors supplied major-index closes, while advancing-issue counts, declining-issue counts, and volume for lesser-followed indexes were harder to obtain. Breadth-and-volume is the missing half of the input: closes alone are incomplete when a model is meant to consume ordered price, volume, or breadth observations.

When a reader wanted high-low range displayed with bar width scaled to volume, the reply directed the reader to packages that support that joint presentation.

What remains unfinished

TradersWeek editorial reading: the three tests are prior to any use of a lookback oscillator as a forecast. They do not score the relative strength index or the commodity channel index. They ask whether session hours, a volatility baseline, and a documented input set have been specified at all.

The archive describes an unfinished starting point. It does not convert that starting point into a finished forecast.