Skip to content
Financial Context Wiki

Relative-Strength Watchlists: A Reproducible Ranking Method

For educational purposes only; not investment advice.

Direct answer

Section titled “Direct answer”

A relative-strength watchlist ranks securities by performance relative to a stated benchmark over fixed lookback windows. For security i, a simple arithmetic score is RSᵢ = total returnᵢ - total returnbenchmark; a wealth ratio such as (1 + returnᵢ) / (1 + returnbenchmark) - 1 is more exact for large moves. Use adjusted prices or total-return data so distributions and splits are treated consistently.

Relative strength is a research filter, not a buy signal. It differs from the Relative Strength Index (RSI), an oscillator calculated from a security’s own gains and losses. Academic momentum is also a broader portfolio effect with specified formation, holding, universe, and implementation rules.

Building a comparable ranking

Section titled “Building a comparable ranking”

Define the investable universe, benchmark, observation dates, return type, currency, and windows before examining results. Windows such as 20, 60, and 120 trading days can show different horizons, but they are conventions rather than natural laws. Compare a stock with both a broad-market benchmark and a relevant industry benchmark to separate market, sector, and company components.

A useful record includes ticker, company, industry, benchmark, 20/60/120-day total returns, excess returns, volume and spread, next filing or earnings date, the fundamental question requiring research, and the date the ranking was calculated. Store raw values and formula version so the list is reproducible.

Ranks can persist because information and investor positioning may adjust gradually, but they also reverse. A high rank says what outperformed, not why, whether the information was already priced, or whether the security is investable at current liquidity and valuation.

Numerical example

Section titled “Numerical example”

Over the same 60 trading days, suppose a broad ETF returns 5%, an industry ETF 18%, and Stock A 35%, all including distributions. Arithmetic industry strength is 18% - 5% = 13 percentage points; stock-versus-industry strength is 35% - 18% = 17 points. The compounded stock-versus-market wealth ratio is 1.35 / 1.05 - 1 = 28.57%.

Stock A enters the research queue; the calculation alone says nothing about future return. Check whether a one-day earnings gap produced most of the move, whether the industry ETF is a fair benchmark, whether estimates and filings support the change, and whether spreads and turnover would materially reduce an implemented result.

Watchlist controls

Section titled “Watchlist controls”
  • Fix the universe at each historical date to avoid survivorship bias.
  • Use data available at the calculation time and lag fundamentals to prevent look-ahead bias.
  • Align dates, market calendars, currencies, corporate actions, and distribution treatment.
  • Separate raw return, market-relative return, and sector-relative return.
  • Flag earnings gaps, mergers, IPOs, stale prices, halts, and insufficient history.
  • Require minimum price, liquidity, and trading history before ranking where appropriate.
  • Record valuation, balance-sheet risk, estimate changes, and the next primary filing to review.
  • Measure turnover, bid-ask spreads, market impact, taxes, and rebalancing frequency.
  • Test multiple periods and out-of-sample dates without choosing the best rule after seeing results.
  • Treat delistings and bankruptcies as outcomes, not missing data.

Common misconceptions

Section titled “Common misconceptions”
  • “Relative strength predicts the next winner.” It summarizes past relative performance.
  • “Relative strength and RSI are the same.” One compares a benchmark; the other is a bounded oscillator.
  • “A strong industry makes every constituent strong.” Weights and company outcomes differ.
  • “Price-only returns are sufficient.” Distributions and corporate actions can change rankings.
  • “The top-ranked stock is undervalued.” Rank contains no valuation conclusion.
  • “Backtested rankings are implementable.” Biases, liquidity, turnover, and costs can erase paper results.
Section titled “Related topics”

Authoritative sources

Section titled “Authoritative sources”