The relative strength index, usually called RSI, is a momentum indicator that scores recent price moves on a scale from 0 to 100. It compares the average size of up moves with the average size of down moves over a set number of periods. A high reading means recent gains have been large relative to recent losses, and a low reading means the opposite. It was introduced by J. Welles Wilder, and it is one of the most widely displayed indicators on trading platforms.
This guide explains how the number is built, shows the arithmetic with invented figures, and sets out how traders use the reference lines and divergences. It also explains why an indicator that looks extreme can stay extreme for a long time. RSI measures the strength of recent moves. It does not measure whether a price is cheap or expensive, and it does not say when a trend will end.
What does RSI measure?
RSI measures momentum, which is how strongly the price has been moving in one direction. It ignores the level of the price and looks only at the changes. A stock that rises steadily in small steps can post a high RSI without moving far from its starting point.
A stock that falls in large jumps can post a low RSI even if it is still expensive by other measures.
The standard setting uses 14 periods. Traders usually see the reading plotted as a line below the price chart, with horizontal reference lines drawn at 70 and 30. A reading above 70 is often described as overbought, and a reading below 30 as oversold. Those labels describe the recent pace of moves, not a verdict on value. Many traders also watch the 50 line as a rough divide between stronger and weaker momentum, although that use is a convention too.
Because the scale is bounded, a reading near either end tells you that moves in one direction have dominated the window. What it cannot tell you is how far the price will travel before the imbalance reverses. That limit is the reason the indicator is best read alongside the trend and the price structure rather than on its own.
How is the number calculated?
RSI is calculated from average gains and average losses over the window. The ratio of the two averages, called RS, is turned into a number between 0 and 100 by the formula 100 minus 100 divided by one plus RS.
Wilder's standard method smooths the averages from one period to the next rather than recalculating them from scratch, so the number changes gradually. The simple version below is shown only to make the arithmetic easy to follow. The figures are invented.
| Input or step | Example A | Example B | What it means |
|---|---|---|---|
| Average gain over the window | 1.0 | 0.5 | Size of the typical up move |
| Average loss over the window | 0.5 | 1.0 | Size of the typical down move |
| RS (average gain ÷ average loss) | 2.0 | 0.5 | Gains relative to losses |
| RSI (100 − 100 ÷ (1 + RS)) | 66.7 | 33.3 | The reading on the 0 to 100 scale |
Example A shows gains twice as large as losses, which produces a reading in the upper part of the scale. Example B shows the reverse. Both examples depend only on the ratio of gains to losses, not on the price level. That is why RSI can be compared across stocks at very different prices, and also why it can miss a move that is large in price terms but balanced in gains and losses.
Why are 70 and 30 used, and why do they fail?
The lines at 70 and 30 are conventions that traders adopted because they often marked turning points in the past. They are a starting point for attention, not a rule for action. A reading that is extreme tells you that momentum is strong, which is different from saying the trend is finished.
In a strong uptrend, RSI can stay above 70 for many periods, and a trader who sells at the first touch will often leave a large part of the move behind. In a strong downtrend, the same is true at 30.
Divergence is the other common reading. Bearish divergence appears when the price makes a higher high while the RSI makes a lower high, which suggests that each new high has less momentum behind it. Bullish divergence is the mirror image at a low. Divergence is easy to spot after the event and harder to use in real time, because the price can keep making new highs for a long time while the indicator stays weak. Treat it as a reason to check the trend and the levels, not as a sell signal by itself.
Divergence also depends on how the highs and lows are chosen. Two traders can mark different swing points on the same chart and reach opposite conclusions. Mark the price extremes first, then read the indicator at those exact points, and write down what counts as a divergence before you look for one.
How do you test an RSI rule?
A useful test begins with a clear rule written in advance: which reading triggers attention, what confirms the signal, and where the trade is wrong. Then run the rule on past charts across many periods and count every signal, including the ones that did not work.
A rule that looks good on a few hand-picked examples often fails when every occurrence is counted.
Keep the settings fixed during the test. Changing the period or the reference lines after seeing the results turns the test into a search for a favourable setting rather than a check of the idea. It is also worth testing across different markets and across periods of both rising and falling prices, because a rule that works only in one kind of market is a narrower finding than it first appears.
A test should also record how often the rule fires. A rule that triggers once a year cannot be judged on a handful of examples, and a rule that triggers every week needs a different kind of scrutiny. The frequency tells you how much evidence the test can produce.
Common mistakes
Selling because RSI is above 70. In a trend, the indicator can stay overbought while the price keeps rising. Check the trend before acting on the reading.
Reading divergence without the price structure. A divergence is one observation. Combine it with levels, the trend and volume before you treat it as meaningful.
Changing the period to suit the chart. A 14-period setting chosen after the fact is not a tested setting. Keep the same length across all your charts.
Checklist
- Confirm the period length and the reference lines on your platform.
- Check the trend on a higher timeframe before you act on an extreme reading.
- Note whether the RSI is above or below 50, which shows the broader momentum.
- Look for divergence only after you have marked the price highs or lows clearly.
- Wait for a confirming signal from the price, such as a break of a level.
- Write down the rule and the invalidation level before you test or trade it.
- Count every signal in your test, including the failures, and judge the rule on the total.
Key takeaways
RSI compares the average size of recent gains with recent losses and scales the result from 0 to 100. The reference lines at 70 and 30 mark extremes of recent momentum, not turning points. Divergence can warn that momentum is fading, but it is easy to read in hindsight. Use RSI as one input, and test any rule on past data before you rely on it.
This guide is for general information, not financial advice. See the disclaimer.