Technical Analysis · Lesson 07 of 15 · Reviewed November 10, 2026
The Relative Strength Index, or RSI, summarizes the balance between upward and downward price changes over a selected calculation history. Its bounded scale makes it easy to display and easy to misunderstand. A high reading does not mean price must fall, and a low reading does not mean price must rebound. The indicator describes a calculation applied to past changes.
This lesson explains a transparent Wilder-style calculation, including the initial averages and later smoothing updates. The dedicated exercise accepts a short chronological close list so beginners can reproduce the arithmetic. It states its flat-series convention explicitly and distinguishes RSI from comparing a stock with a benchmark.
1. Start with changes, not price levels
RSI uses changes between consecutive observations. For closes 50, 51, 50, 52, 51, 53, the changes are +1, −1, +2, −1, and +2. Each change contributes either a nonnegative gain or a nonnegative loss magnitude. A decline of $1 enters the loss series as positive 1, not negative 1.
For those five changes, gains are 1, 0, 2, 0, 2 and losses are 0, 1, 0, 1, 0. The sum of gains is 5 and the sum of loss magnitudes is 2. A five-change initial average therefore gives average gain 1 and average loss 0.4. At least six closes are needed to obtain five changes.
This requirement is easy to miss. A period counts changes in this worksheet, so the minimum close count is period plus one. Supplying five closes for a five-period initial calculation leaves only four changes. The tool should reject that insufficient history rather than silently shortening the period.
2. Convert the averages to an index
Relative strength in the RSI formula is average gain divided by average loss magnitude. In the example, it is 1 ÷ 0.4 = 2.5. RSI equals 100 minus 100 divided by one plus that ratio. The result is approximately 71.43. An equivalent expression is 100 times average gain divided by average gain plus average loss when the denominator is positive.
The result is an index value, not a percentage probability of an advance. RSI 71.43 does not mean a 71.43% chance of a winning trade. It also does not mean the stock gained 71.43% over the sample. Those interpretations use different concepts and denominators.
Price scale alone does not create the same effect as momentum. Multiplying every supplied close by a positive constant multiplies gains and losses together and leaves their ratio unchanged. That mathematical property does not make RSI equally useful across all instruments, intervals, or market conditions.
3. Apply Wilder-style smoothing to later changes
After the initial averages, this worksheet updates average gain as prior average gain times period minus one, plus the new gain, all divided by period. Average loss is updated in the same way using the new loss magnitude. This recursive process carries forward earlier history rather than discarding all observations outside a simple rolling window.
Add a new close of $52 after the example’s final $53. The new change is −$1. With period five, the updated average gain is (1 × 4 + 0) ÷ 5 = 0.8. Updated average loss is (0.4 × 4 + 1) ÷ 5 = 0.52. RSI becomes approximately 60.61. The new decline changes the balance without replacing the calculation with a fresh five-change arithmetic mean.
Initialization matters. A platform using a longer earlier history can differ from a worksheet seeded at the first supplied observation. The exercise is intended to teach the calculation, not guarantee exact agreement with every platform’s historical state. Record the seed, history length, interval, and source price when reconciling values.
4. Handle zero-gain and zero-loss cases explicitly
If average loss is zero and average gain is positive, the worksheet reports RSI 100. If average gain is zero and average loss is positive, it reports RSI 0. These endpoint cases describe one-sided changes under the supplied history; they do not establish an inevitable reversal.
If both averages are zero, the gain-loss ratio is mathematically indeterminate. This worksheet uses a neutral display convention of RSI 50 for an entirely flat state and labels that convention. Other implementations can handle such cases differently. A stated convention is preferable to a hidden division error or unexplained empty value.
A flat price series can also indicate sparse data or stale observations. The numerical convention does not establish that the market was balanced or liquid. Inspect the data source before interpreting a long sequence of unchanged closes as a meaningful momentum condition.
5. Interpret familiar thresholds as descriptions
Levels around 70 and 30 are commonly discussed as overbought and oversold reference areas. Those names can encourage an unhelpful conclusion that high must mean sell and low must mean buy. A strong trend can retain elevated or depressed readings for an extended period, and a threshold alone does not supply an entry, stop, or horizon.
A more precise statement is “the selected RSI calculation ended above the chosen upper reference.” If a study uses that event, define whether it requires a crossing, a close, a duration, or a return through the threshold. These are different conditions and can yield different samples.
The relationship to wider structure matters. A pullback within an upward sequence and an accelerating decline can both produce a low short-period RSI. The trend structure lesson helps keep the momentum calculation separate from the larger selected price sequence.
6. Understand divergence without treating it as certainty
A divergence compares selected price extremes with selected indicator extremes. For example, price might make a higher selected high while RSI makes a lower selected high. That can be framed as a question about changing momentum, but the result depends on which pivots were selected and when they became confirmed.
A historical chart can make divergence obvious only after later bars identify the turning points. Preserve the confirmation delay. A study that labels the final high immediately using future observations creates the same look-ahead problem encountered in pivot analysis.
Divergence can persist while price continues in the original direction. It should not be presented as a countdown to a reversal. A testable method needs a separate timing rule, failure condition, execution assumption, and sample review. The indicator’s visual appeal does not remove those requirements.
7. Use the dedicated RSI exercise
Open the Lesson 7 RSI exercise. Enter closes 50, 51, 50, 52, 51, 53 and period 5. The tool should show the seeded average gain of 1, average loss of 0.4, and RSI near 71.43. The short period is chosen for hand calculation, not recommended as a trading setting.
Append 52 and calculate again. The recursive update should produce average gain 0.8, average loss 0.52, and RSI near 60.61. Compare that result with the arithmetic above. The visual series begins only when the initial period is available and then updates with each additional supplied close.
Test an increasing list, a decreasing list, and an unchanged list. Predict the endpoint conventions before calculating. Finally supply too few closes. The tool should explain the minimum requirement clearly. These checks make the formula and its edge cases easier to understand than memorizing a colored oscillator panel.
8. Distinguish RSI from benchmark relative strength
Despite the name, this RSI calculation does not divide a stock’s performance by a market index. It compares upward and downward changes within the same price series. Benchmark relative strength is a different analysis that compares two instruments over aligned dates and consistent return conventions.
A stock can have a high RSI and still underperform a benchmark that rose faster. It can also outperform a falling benchmark while losing money in absolute terms. These observations are compatible because the measurements answer different questions.
The course’s relative strength and benchmark lesson provides a separate exercise for rebasing two series. Keeping the two concepts distinct prevents a common misunderstanding caused by their similar names.
9. Design a momentum study carefully
Specify the period, input price, initialization, threshold rule, completed-bar requirement, observation horizon, and eligible instruments. Record how equal or missing observations are handled. A screenshot labeled “RSI signal” omits too much information to support a reproducible comparison.
If several thresholds or periods are tested, retain the search history and use genuinely unseen observations for later evaluation. Selecting the best-looking combination and reporting only its result overstates what the exercise established. Transaction costs and repeated reversals can materially change a strategy built from frequent crossings.
Consider whether RSI adds distinct information to other conditions. Combining it with several price-derived momentum indicators can count similar evidence repeatedly. The MACD lesson describes another momentum summary and explains why different displays do not necessarily create independent confirmation.
10. Check that the interpretation matches the arithmetic
You should be able to separate gain and loss magnitudes, seed the initial averages, apply a later Wilder update, and explain all three zero cases. You should also be able to explain why RSI is neither a probability nor a benchmark comparison.
A responsible RSI note states the calculation and a narrow observation before proposing a hypothesis. The goal is to understand what changed in the supplied series and how the chosen smoothing represents it. That understanding supports careful experimentation without turning a familiar oscillator into an unsupported instruction to trade.
Case review: the seed and the later update
With period five and closes 50, 51, 50, 52, 51, 53, the initial gain and loss averages are 1 and 0.4. Add a close of 52. The next averages become 0.8 and 0.52, producing RSI about 60.61. A fresh arithmetic average of only the latest five changes would answer a different calculation question.
Now imagine a platform began its recursive history several months earlier. Its current averages can differ from this short seeded worksheet even when both use the same update formula. The discrepancy should trigger a definition check rather than an immediate claim that one source is wrong.
For the journal, preserve the complete supplied list and identify where the initial seed occurs. Add one close at a time and record how the state changes. This exercise teaches the recursive memory of the indicator and makes initialization visible. It also discourages treating a displayed threshold as independent of the underlying calculation history. The same final price can coexist with different RSI values when the preceding path differs.
A repeatable practice and review protocol
Before changing the example, write down the question you expect the calculation to answer. Copy the input values, units, chronological order where relevant, and the observation time. State whether the values are fictional, historical, or current. The exercises in this course begin with fictional data so the arithmetic can be studied without implying that a real security offers a particular opportunity. A calculator result is only as meaningful as its input definitions.
Predict the direction of one change before pressing the calculation button. Change one input at a time, explain which part of the formula it affects, and compare the result with your prediction. This controlled approach is more informative than changing several fields until an attractive number appears. If the result surprises you, return to the definition and denominator before looking for a market explanation. Arithmetic mistakes and interpretation mistakes require different corrections.
Then test an ordinary alternative and an edge case. An ordinary alternative checks whether you understand the relationship away from the default example. An edge case checks whether a ratio becomes undefined, a sample becomes insufficient, or an input ordering becomes impossible. A clear validation message is a useful result. Do not work around it by entering an unrelated value merely to make the screen produce an answer. First understand which assumption the rejected value violated.
Save a short review with four parts: the observation, the calculation, the interpretation, and the unresolved limitation. Keep those parts separate. The observation records what the supplied data says; the calculation shows how the number was obtained; the interpretation is a conditional hypothesis; the limitation identifies information the exercise does not provide. This structure makes the note understandable to another beginner and makes later revisions easier to audit.
If you extend the example into a paper-trading study, define signal time, decision time, order assumptions, quantity, costs, and review horizon before checking the outcome. Preserve the original version. Record unfavorable cases and skipped cases under the same rules as favorable ones. A chart image can support the record, but it should not replace the actual values and definitions. A successful-looking historical picture is not a substitute for a complete sample.
Finish by explaining the result in plain language without using a pattern name as the explanation. Identify what changed, what stayed fixed, and what the exercise still cannot establish. You have completed the practice when you can reproduce the important arithmetic, recognize an invalid comparison, and describe an alternative outcome. Repeating this process across the course builds a consistent analytical habit while keeping the different formulas and their limitations distinct.
Sources and calculation notes
Primary educational background provides additional context for the topic. The fictional examples, teaching cases, diagrams, and worksheets here are original educational illustrations. Calculation conventions and missing information are stated explicitly; indicator readings do not establish future returns.