How to Use RSI for Overbought/Oversold Signals

What RSI Actually Measures

The Relative Strength Index (RSI) is a momentum oscillator, developed by J. Welles Wilder Jr. and introduced in his 1978 book New Concepts in Technical Trading Systems. It measures the speed and magnitude of recent price movements on a scale of 0 to 100 — essentially a speedometer for momentum, not a predictor of exact turning points.

The RSI Formula

RSI = 100 − (100 ÷ (1 + RS))

Where RS (Relative Strength) is the average gain over a set period — typically 14 candles — divided by the average loss over that same period. You don't need to calculate this by hand; every charting platform plots it automatically. What matters is understanding what drives the number: strong, consistent gains push RSI toward 100, while strong, consistent losses push it toward 0.

Reading Overbought and Oversold

Using the standard 14-period setting, an RSI reading above 70 is conventionally labeled overbought, suggesting price has risen quickly and may be due for a pause or pullback. A reading below 30 is labeled oversold, suggesting the opposite. These thresholds work reasonably well in range-bound markets, where price oscillates between clear boundaries rather than trending strongly in one direction.

Why Overbought Isn't a Sell Signal

This is the single most important nuance in using RSI correctly, and it's where most beginners get burned. In a strong trending market, RSI can remain above 70 — or below 30 — for an extended stretch, sometimes weeks, while price simply continues in the same direction. Traders who treat "overbought" as an automatic sell signal during a genuine strong trend can find themselves shorting a market that keeps climbing for considerably longer than expected.

The label "overbought" describes momentum, not an expiration date on the trend. It's a caution flag worth pairing with other confirmation — not a standalone trigger to trade against a strong trend.

Adaptive Thresholds for Trending Markets

Because static 70/30 levels fire too early in strong trending conditions, many experienced traders shift to 80/20 thresholds on trending instruments — commonly cited for assets like gold and major indices — to reduce false signals. In a strong uptrend, "oversold" might realistically show up closer to 40-45 rather than the textbook 30; in a strong downtrend, "overbought" might realistically sit closer to 55-60.

The practical takeaway: match your thresholds to the actual behavior of the instrument and current market regime, rather than applying the same static numbers to every chart regardless of conditions.

RSI Divergence: The Highest-Value Signal

Divergence occurs when price and RSI move in opposite directions — and many experienced traders consider it the single highest-value RSI signal, more reliable than a simple threshold cross.

• Regular bullish divergence — price makes a lower low, but RSI makes a higher low. This suggests downward momentum is weakening even as price continues falling, often preceding a reversal.
• Regular bearish divergence — price makes a higher high, but RSI makes a lower high, suggesting upward momentum is fading even as price pushes higher.
• Hidden divergence — the inverse pattern, which typically signals trend continuation rather than reversal, and is often used by trend-following traders looking to add to an existing position.

When trading divergence, the 70/30 threshold lines matter far less than comparing the swing highs and swing lows RSI is printing against the swing highs and lows on the price chart itself.

Divergence is most meaningful when it forms near a real level — see Support and Resistance: The Foundation of Price Action.

Choosing Your RSI Settings

• 14-period is the standard default, and a reasonable starting point across most timeframes and markets.
• Shorter periods (such as 7) react faster and suit shorter-term day trading, at the cost of more noise.
• Longer periods (such as 21) smooth out short-term fluctuations, better suited to swing trading.

A sensible approach: stick with the 14-period default and standard 70/30 thresholds across a meaningful number of trades before adjusting anything — changing settings too early makes it difficult to tell whether a strategy has a genuine edge or you're just curve-fitting to recent price action.

A Real Divergence Example

Say gold makes a new high at $2,350, then pulls back, then rallies again to a slightly higher high at $2,360. On the surface, this looks like continued strength. But if RSI made a lower high on that second peak compared to the first, that's bearish divergence — price is making new highs, but the underlying momentum driving those highs is actually weakening.

This doesn't guarantee an immediate reversal, but it's a genuine warning sign worth taking seriously, especially when it appears at a level that already has other reasons to expect resistance (a prior high, a round number, a trendline). Divergence is best treated as one meaningful piece of evidence among several, not a standalone trading signal.

Frequently Asked Questions

What is a good RSI level to buy or sell?

The traditional thresholds are RSI above 70 for overbought and below 30 for oversold, set by J. Welles Wilder in 1978 using a 14-period calculation. In strong trending markets, many traders adjust these to 80/20 to reduce false signals, since RSI can remain above 70 or below 30 for extended periods during a strong trend.

Is overbought RSI always a sell signal?

No. In a strong uptrend, RSI can remain above 70 for weeks while price continues climbing. Treating overbought as an automatic sell signal during a strong trend is one of the most common RSI mistakes and can lead to shorting a market that keeps rising.

What is RSI divergence?

RSI divergence occurs when price and the RSI indicator move in opposite directions. Regular bullish divergence happens when price makes a lower low but RSI makes a higher low, suggesting weakening downward momentum. Regular bearish divergence is the opposite — price makes a higher high while RSI makes a lower high.

What is the RSI formula?

RSI = 100 − (100 ÷ (1 + RS)), where RS is the average gain over a set period (typically 14) divided by the average loss over that same period. The result is a value between 0 and 100.