Methodology / Momentum

Momentum: the shortest memory, and the least weight

Momentum describes how quickly price has moved recently. It is the shortest-memory factor in the set, the noisiest, and of the six it is the one the composite leans on least — deliberately.

One of the six factors in the Tapeline composite. Counts for less than any other factor in the composite. Tapeline publishes the ordering of the factor weights, not the numeric weights or the scoring equation.

What it measures

Observable quantities, nothing else

A momentum-quality reading for the ticker, supplied by the upstream signal system either as a number or as a graded label.
A short-horizon return, approximated by rescaling the ticker's multi-month price change.

Where this falls short

The short-horizon component is an approximation, derived by rescaling a longer multi-month return rather than measured over a short window directly. It is therefore smoother than a true short-window reading and will lag a genuine recent turn. This is the weakest-constructed of the six factors, and it carries the least weight for that reason.

How the reading is derived

The same procedure on every ticker

01Each available component is mapped onto a common 0–100 scale and the available ones are averaged.
02Where the momentum-quality input arrives as a label rather than a number, the label is resolved to a value on the same scale.
03The composite weights Momentum least of the six factors, because short-horizon rate of change reverses often. The ordering of the factor weights is published on /how-it-works; the numeric weights are not.

When the data is missing

When a factor cannot be computed for a ticker, the composite substitutes a mid-range value rather than a zero, so a missing input does not drag the score down. The trade-off is that a mid-range reading can mean 'measured, and unremarkable' or 'not available' — the per-ticker confidence percentage is what separates the two.

Tapeline deliberately does not publish the numeric weights, the scoring equation, or the exact band edges used to map a measurement onto the 0–100 scale. What is published is the factor set, the weight ordering, each factor’s contribution on every ticker, and the record of every daily top-10.

What feeds it

Data behind this factor

Live market data

Daily and intraday bars. Sub-60 seconds during US market hours.

Upstream signal system

The momentum-quality reading, supplied per ticker by Tapeline's scoring pipeline and refreshed on the worker tick.

Every category, its refresh cadence and where it appears in the product is listed on data sources.

Known limitations

Where this factor is weak

Every one of these is a property of the method, not a bug waiting to be fixed. They are listed here so a reader can decide how much weight to give the reading.

The short-horizon input is a proxy, not a measurement. A sharp move in the last few weeks reaches this factor diluted by the longer window it is derived from.
It is sensitive to one-off events — an earnings gap, a news spike, an index-inclusion flow — and cannot distinguish those from a sustained move.
Low-float and low-liquidity tickers produce large readings from small dollar flows. Tapeline applies a liquidity floor to the ranked scanner and the public scorecard for this reason; that floor can be switched off on the scanner to browse the full scored universe.
Where the upstream momentum-quality input is a coarse label rather than a number, the reading it produces is correspondingly coarse.
It carries no information about the company at all.

Limits that apply to the whole product, rather than to this factor, are on limitations.

Common questions

Momentum FAQ

Why is Momentum weighted least?

Two reasons, both stated plainly: short-horizon rate of change reverses often enough to be a noisy input on its own, and one of its two components is an approximation rather than a direct measurement. The relative ordering of the six factor weights is published; the exact numeric weights are not.

Why does the Momentum reading change less than I expect day to day?

Because its short-horizon component is derived by rescaling a longer multi-month return rather than measured over a short window. That makes it smoother, and slower to reflect a genuine recent turn, than the factor's name suggests.

The other five