Relative Strength vs RSI: Two Different Measures
Ben Ghabili · Published
Relative strength and the Relative Strength Index sound interchangeable. They are not. One compares an asset with something else; the other examines changes in the asset's own price.
Before interpreting either, identify the calculation. A label containing “strength” does not tell you whether a stock rose, beat a benchmark or has a particular momentum reading.
What is the difference between relative strength and RSI?
Comparative relative strength measures an asset's performance against another asset or benchmark. The Relative Strength Index, or RSI, is a momentum oscillator based on upward and downward changes in one asset's own price. Ordinary RSI has no benchmark input, so a high RSI does not establish benchmark outperformance.
| Measure | Comparison | Key question | Important limit |
|---|---|---|---|
| Comparative relative strength | Asset versus selected benchmark | Which performed better over the defined window? | Outperformance can occur while both fall |
| RSI | Upward versus downward changes in the asset's own price | What does the selected calculation show about its price momentum? | It does not compare the asset with the market |
The distinction matters when reviewing rankings or charts. Always check the metric's definition rather than assuming every “RS” field uses the same calculation.
How is comparative relative strength calculated?
A simple relative-strength line divides the asset's price by the benchmark's price at matching observations. A rising ratio indicates improving relative price performance over that interval; it does not require the asset's own price to rise.
Relative-price ratio = asset price ÷ benchmark price.
The raw ratio's level depends on the price units and starting levels. Focus on its change, with comparable data, rather than treating a ratio above one as inherently stronger than a ratio below one.
Return comparisons can also be expressed as a difference in percentage points. That is a different calculation from the percentage change in a price ratio. Both can be useful, but they should not share an unexplained label.
Can relative strength improve while a stock falls?
Yes. A stock improves relative to a benchmark if it loses less over the same period. Relative outperformance describes a comparison; absolute return describes what happened to the asset itself.
Consider two fictional period-end daily observations, with both series in the same currency. Inputs are invented and exclude dividends, costs, currency changes and corporate actions.
| Measure | Start | End | Period return |
|---|---|---|---|
| Asset price | 100 | 90 | −10% |
| Benchmark price | 100 | 80 | −20% |
| Asset price divided by benchmark price | 1.000 | 1.125 | +12.5% ratio change |
The asset fell 10% and the benchmark fell 20%. The relative-price ratio rose 12.5%, because 1.125 ÷ 1.000 − 1 = 0.125.
The difference in their returns is 10 percentage points: −10% − (−20%). That is not a 12.5-percentage-point gap. The ratio change and return difference use different denominators.
The asset outperformed in this example, but an unhedged holder still suffered a price loss before costs. Neither relative result establishes that the asset was a good purchase or will lead in the next period.
What does RSI actually measure?
RSI transforms average upward and downward price changes into a reading between zero and 100. A higher reading reflects relatively stronger upward changes under the selected calculation. It is not a percentage investment return or a probability of the next move.
For a hypothetical formula exercise, assume a 14-daily-session calculation has an average upward price change of $1.00 and an average downward change magnitude of $0.50. These invented averages are formula inputs, not observations claimed from a real chart.
Their ratio is 2. The basic RSI transformation gives:
RSI = 100 − 100 ÷ (1 + 2) ≈ 66.67.
No benchmark appears. You could change the benchmark's performance without changing these asset-price inputs, and this RSI calculation would remain 66.67.
The averages must follow the implementation's stated convention. Period length, observation frequency and averaging or smoothing method need to be consistent when comparing readings. Do not assume a value from one setting is equivalent to a value from another.
Does an overbought RSI mean you should sell?
No. “Overbought” is a conventional indicator description, not an instruction or proof that a reversal is imminent. Readings above 70 and below 30 are traditional overbought and oversold thresholds, but RSI can remain in those regions during strong trends.
A high RSI does not establish that a stock is fundamentally expensive. A low RSI does not establish undervaluation. Those conclusions require business and valuation evidence outside the oscillator.
If you use an RSI condition in a trading process, define the rule and evaluate its actual outcomes. Naming a familiar threshold does not demonstrate an advantage after costs.
How should you compare strength across timeframes?
Compare like with like. For comparative strength, specify the benchmark, start and end observations, currency and price-versus-total-return basis. For RSI, specify the asset, observation frequency, period and averaging convention.
A positive three-month relative result and a weaker one-month result can coexist. They use different starting points. A short-term RSI change does not settle which asset led over three months.
Use Tapelab's sector page to inspect sector performance across its displayed windows and its three-month comparison with SPY. The page labels its ETF returns as trailing total returns and its relative-strength measure as the change in the ETF's price ratio to SPY. Keep those different bases and the selected dates visible when comparing columns. The relative comparison should not be mistaken for RSI.
Before using “strong” in a research note, finish the sentence: strong relative to what, measured how, and over which period? Then record the absolute return separately. That keeps leadership, momentum and actual gains from becoming one ambiguous claim.
Nothing here is investment advice.