Relative strength against a benchmark
What it measures¶
Relative strength compares a share with the market rather than with its own past. A share that rose 10% while the market rose 20% has lagged, even though its own return was positive. The comparison uses a benchmark — for shares listed in the US, SPY, an exchange-traded fund (a fund whose units trade like a share) that tracks the S&P 500 index of 500 large US companies.
Two figures appear on the pages:
- the ratio line — the share's price divided by the benchmark's, week by week: rising means doing better than the market, whichever way the market itself moves;
- Mansfield relative strength — that ratio against its own 52-week average: (ratio ÷ 52-week average of the ratio − 1). Above zero, the share has done better than the market lately compared with its past year; below, worse. See Mansfield relative strength.
On the Price & timing strip this tile has one vote, R1: is Mansfield relative strength above zero on weekly bars? With a single vote it prints a noun instead of a verdict word: leading (above +1%), in line (within ±1%) or lagging (below −1%).
Why it is on the page¶
Stan Weinstein treated a share's strength against the market as a confirmation of its own trend: an advance that also beats the market is broader-based. On the Technical page the same readings sit in the relative-strength panel, with the share's and the benchmark's weighted returns (RSraw).
Thresholds, their source and the population they were fitted on¶
| Mansfield reading | Noun | Vote |
|---|---|---|
| above +1% | leading | ✓ passed |
| 0 to +1% | in line | ✓ passed |
| −1% to 0 | in line | ✕ failed |
| below −1% | lagging | ✕ failed |
The vote counts in the strip's n of m votes passed; the tile itself shows the noun and no verdict word.
- Source: Stan Weinstein's zero line; the Mansfield formula as published by stage-analysis practitioners.
- Fitted on: no statistical sample for the formula; the momentum research the relative idea rests on is peer-reviewed and replicated.
- Evidence grade: practitioner convention (the formula); peer-reviewed and replicated (the concept).
A 1–99 "relative strength rating" of the kind some publishers print needs a ranked universe of thousands of shares; the product's catalogue is far smaller, so no such rank is ever shown.
A worked example¶
On weekly bars to 23 October 2026, against SPY:
- Apple — Mansfield relative strength +13.2%: leading, vote passed.
- Honeywell International — −8.1%: lagging, vote failed.
What “No verdict” means here¶
The tile needs a stored benchmark series and at least 52 weekly points of the ratio. Without a benchmark it reads No verdict — benchmark series not stored; with too short a history, No verdict — fewer than 52 weekly ratio points. Either way its vote leaves the strip's count. For the benchmark itself — SPY — the comparison is not applicable: a share cannot be compared with itself.
What the verdict is and is not¶
The noun describes how the share has moved against the market over the past year. It is not a forecast that the gap will persist, not a ranking among companies, and not an instruction to trade the shares.