RSI divergence is when price and the RSI disagree about a swing. Price makes a higher high but the RSI makes a lower high (bearish divergence), or price makes a lower low while the RSI makes a higher low (bullish divergence). It tells you the latest push had less momentum behind it than the one before. It doesn't tell you the move is over. In a strong trend, divergence can show up three or four times before anything turns. It fails often, and nothing on this page is financial advice.
Why RSI divergence happens
Divergence is arithmetic. J. Welles Wilder's RSI, from his 1978 book New Concepts in Technical Trading Systems, compares the average size of up closes with the average size of down closes over the last 14 periods: RS is average gain divided by average loss, and RSI = 100 minus 100 / (1 + RS). (StockCharts' ChartSchool page walks through the full calculation.)
So RSI doesn't care how high price is. It cares how price got there. In the drawing above, the first high came at the end of five strong up candles in a row and the RSI read 87.7. The second high was 0.7 points higher, but it was reached by stepping up and down, with small gains and frequent losses. RSI on that bar: 72.5. Higher price, weaker push.
| Avg gain / loss | RSI | |
|---|---|---|
| First rally | 0.60 / 0.10 | 85.7 |
| Second rally | 0.40 / 0.15 | 72.7 |
Illustrative averages. RS is 6.0 on the first rally and about 2.7 on the second, which sits at a higher price. A smaller average gain or a bigger average loss pulls RSI down even while price climbs.
The four types of RSI divergence
| Type | Price / RSI | Hints at |
|---|---|---|
| Regular bullish | Lower low / higher low | Selling fading after a decline |
| Regular bearish | Higher high / lower high | Buying fading after a rally |
| Hidden bullish | Higher low / lower low | An uptrend resuming after a pullback |
| Hidden bearish | Lower high / higher high | A downtrend resuming after a bounce |
The quick way to remember it: in a regular divergence, price makes the new extreme and RSI refuses to follow, which hints at a turn. In a hidden divergence, RSI makes the new extreme and price refuses, which hints at continuation. Hidden divergence is close to what Andrew Cardwell, Constance Brown's RSI mentor, called positive and negative reversals, which StockCharts describes as the opposite of bearish and bullish divergences.
How to draw RSI divergence without fooling yourself
- 1
Find two clear swings on the price chart
Two highs for a bearish divergence, two lows for a bullish one. They should stand out from the candles around them. If you have to squint to see a swing, it's not one.
- 2
Use the same bars on the RSI
Drop a vertical line from each price swing to the RSI panel and read the RSI there. Don't pick a different RSI peak two bars away because it makes a nicer line.
- 3
Stay consistent with closes or wicks
RSI is built from closes, so joining closes on the price chart is the cleaner comparison. If you prefer wicks, use wicks every time. Mixing them invents divergences.
- 4
Prefer a first swing beyond 70 or 30
StockCharts' ChartSchool notes that divergences carry more weight when they form after an overbought or oversold reading. A divergence that starts from an RSI of 55 says very little.
- 5
Wait for price to confirm
The divergence is a condition, not a trigger. Wait for price to break the small swing between the two highs or lows, or for a clear reversal candle such as a bullish engulfing candle.
Where traders put the entry, stop and target
Here price falls to 94.4 with the RSI at 12.3, deeply oversold. It bounces, then sinks to a lower low of 94.0, but RSI only gets down to 25.4. The second wave of selling was weaker. Nothing to do yet.
The trigger comes when price closes above the bounce high between the two lows (95.80). A common plan, using these made-up prices:
- Entry above the bounce high, 95.85.
- Stop just under the second low, 93.79. If price makes a new low beyond that, the divergence has failed, simple as that.
- Target at the last small swing high before the slide, 98.60.
That's 2.06 of risk for 2.75 of reward, about 1 to 1.3. Waiting for the break costs you part of the move, and that's the price of skipping the divergences that never turn. The stop distance also sets the position size, which the position size calculator works out.
When divergence fails: the strong trend
This is the chart that costs divergence traders the most. In a steady uptrend, each new high comes with a little less momentum than the last, so RSI peaks drift lower: 87.2, 80.8, 75.8, 72.2 in the drawing. Each pair is a textbook bearish divergence. Someone shorting each one is stopped out three times while price grinds higher.
StockCharts makes the same point with a real chart of the SPDR S&P 500 ETF showing three bearish divergences and a continuing uptrend, and warns that divergences "are misleading in a strong trend". Constance Brown's work, also summarised there, adds a useful filter: in bull markets RSI tends to hold a 40 to 90 range, and in bear markets 10 to 60. If RSI keeps bouncing off 40 to 50 in an uptrend, the trend is healthy, divergence or not.
If you take one rule from this page: do not trade a regular divergence against a strong trend unless price has also broken its structure. A failed higher low in an uptrend, or the neckline of a double top, says far more than the RSI line does.
Hidden divergence, in practice
Hidden divergence is most useful as a filter for trades you were already planning. In an uptrend, price pulls back to a higher low while RSI dips to a lower low than on the last pullback: momentum reset while the structure held. StockCharts adds that the RSI low in a positive reversal usually sits between 30 and 50, not in oversold territory. If that pullback also lands on support or a moving average you trust, the hidden divergence is one more reason to look for a long. On its own, without a trend to continue, it's not a setup.
What about the RSI divergence win rates online?
In one search we ran in November 2026, the top result for "rsi divergence trading strategy" was an RSI strategy page promising a win rate above 90 percent in its title. A win rate on its own tells you very little. You also need the average win, the average loss, the market, the period tested and the costs. Plain arithmetic shows why:
| Strategy | Avg win / loss | Per trade |
|---|---|---|
| A, wins 90% | 0.10R / 1R | −0.01R |
| B, wins 40% | 2R / 1R | +0.20R |
Made-up strategies. R is the amount risked per trade. A: 0.9 × 0.10 minus 0.1 × 1 = −0.01R. B: 0.4 × 2 minus 0.6 × 1 = +0.20R. Strategy A wins nine times in ten and still loses money.
We won't quote a win rate for RSI divergence on this page, because no honest single number exists: it changes with the market, the timeframe, the swing rules and the exit. If you test it yourself, log every signal including the ones you skip, and judge it on average R. The risk reward calculator shows the win rate each ratio needs just to break even.
Reading RSI on your own chart with WolfGPT
Upload a chart screenshot to WolfGPT and the read includes an RSI reading with an overbought, neutral or oversold label, next to the trend, the support and resistance levels and the candles it's looking at. It doesn't draw divergence lines for you. Use the read to check the levels and the trend, then compare the two swings yourself with the steps above. For the price structures that give divergence its trigger, see the chart patterns guide.
Before you trade an RSI divergence
- Two clear swings on price, and the RSI read on exactly the same bars.
- The first RSI swing was beyond 70 (bearish) or 30 (bullish).
- You checked the bigger trend. A regular divergence against a strong trend needs a break in price structure too.
- Price has confirmed: a close beyond the swing between the two highs or lows, or a clear reversal candle.
- The stop sits beyond the second swing, and the target gives a reward worth that risk.
- Any divergence an indicator flagged has been checked by eye on closed candles.
Questions people ask
What RSI setting is best for divergence?
Most traders use Wilder's default of 14 periods. A shorter setting such as 7 swings harder and produces many more divergences, most of them noise; a longer one such as 21 gives fewer and later. Pick one and keep it, so your swings are comparable from chart to chart.
Which timeframe works best for RSI divergence?
Divergence on higher timeframes (4-hour, daily, weekly) reflects more trading and tends to matter more. On 1-minute and 5-minute charts it appears constantly. A common approach is to find the divergence on a higher timeframe and look for the confirming break on a lower one.
Is RSI divergence the same as MACD divergence?
Same idea, different indicator. MACD is built from moving averages of price, so its divergences often appear at the same swings but can disagree with RSI on smaller ones. When both diverge at a major level, traders take it more seriously, but two indicators built from the same prices aren't two independent opinions.
Does RSI divergence work on crypto and forex?
The calculation works on any market with closing prices. The weaknesses travel too: crypto trends can run hard for weeks with repeated divergence, and forex pairs can grind in ranges where divergence appears at every swing. Judge it per market from your own log, not from someone else's win rate.
What is an RSI failure swing?
Wilder's other signal, which ignores price entirely. A bullish failure swing is RSI dropping below 30, bouncing, pulling back while holding above 30, then breaking its bounce high. The bearish version is the mirror image above 70.
Can a scanner find RSI divergence for me?
Yes, scanners and indicators that flag divergence exist on most platforms, including stock screeners popular in India. They use fixed rules to pick swings, so they miss some and flag others late. Use them to build a shortlist, then check each chart by eye.