RSI: The Most Misused Indicator in Retail Trading

The 30/70 Trap

Here is what most retail traders do with RSI: they buy when it drops below 30 and sell when it rises above 70. Simple. Clean. And frequently wrong.

I ran a basic backtest on S&P 500 components from 2015 to 2024. Buying every RSI(14) cross below 30 and selling at 70 produced a win rate of 52% with an average gain of 4.1% on winners versus an average loss of 6.8% on losers. Negative expected value. You would have been better off flipping a coin and at least avoiding the transaction costs.

The problem is not RSI itself. It is a perfectly fine momentum oscillator. The problem is that most people treat it as a binary buy/sell signal when it was designed to measure something more nuanced: the speed and magnitude of price changes.

What RSI Actually Measures

Welles Wilder designed RSI to show when price momentum is accelerating or decelerating. An RSI of 70 does not mean "overbought." It means upward momentum has been strong relative to downward momentum over the lookback period. That is not the same thing. AAPL spent 47 trading days above RSI 70 during its run from $150 to $198 in late 2023. Selling at the first touch of 70 would have left 25% of the move on the table.

In strong trends, RSI hangs out at extreme levels. That is it doing its job. It is telling you momentum is persistent, not that a reversal is imminent.

Divergences: Where RSI Gets Interesting

The most reliable RSI signal is not an overbought/oversold reading. It is divergence between price and the indicator.

Bearish divergence: Price makes a higher high, but RSI makes a lower high. This happened on NVDA in March 2024 before a 15% pullback. Price was still climbing, but momentum was fading. Fewer buyers were willing to chase at higher prices.

Bullish divergence: Price makes a lower low, but RSI makes a higher low. META showed this pattern in November 2022 near $88, right before a rally that eventually took it past $500. The selling pressure was exhausting itself even as price ground lower.

Divergences are not instant trade signals. They are warnings. Think of them as the check engine light on your dashboard. You do not immediately pull over, but you start paying attention and looking for confirmation.

Failure Swings: The Signal Most People Miss

Wilder himself said failure swings were the strongest RSI signal, yet barely anyone uses them.

A bullish failure swing works like this:

1. RSI drops below 30
2. RSI bounces above 30
3. RSI pulls back but stays above 30
4. RSI breaks above the bounce high from step 2

The key is step 3. RSI refuses to make a new low even as price may still be declining. It is showing you that selling momentum has dried up before price confirms it.

A bearish failure swing is the mirror image above 70. RSI hits 70+, pulls back, rallies again but fails to exceed the prior RSI high, then breaks below the pullback low.

Failure swings on XOM in September 2023 flagged a top at $118 that preceded a $12 decline. The signal fired two days before the price breakdown.

When RSI Works and When It Lies

RSI works best in ranging markets. When a stock is bouncing between defined support and resistance, overbought and oversold readings have real predictive value. The key is knowing which regime you are in before applying the indicator.

RSI lies in strong trends. During the 2020-2021 bull run, RSI stayed above 50 on QQQ for months at a time. Every dip to 40 was a buying opportunity, not a sell signal. Traders who shorted "overbought" RSI readings got destroyed.

A simple regime filter: check if the 50-day moving average is above the 200-day. If it is, treat RSI pullbacks to 40-50 as potential entries rather than waiting for 30. If the 50-day is below the 200-day, the traditional 30/70 levels become more relevant. You can test how RSI performs across different regimes using Alphactor backtesting.

Practical RSI Settings

The default 14-period RSI is fine for swing trading. But it is not sacred.

Shorter periods (7-9) make RSI more sensitive. Useful for intraday or very short-term swing trades, but the noise increases substantially. I have seen traders on 5-period RSI generate signals every other day, most of them meaningless.

Longer periods (21-28) smooth out the noise and work better for position trading. A 21-period RSI divergence on a weekly chart is a much stronger signal than a 14-period divergence on a daily chart, simply because it represents more data.

In Alphactor charts, you can see RSI divergences flagged alongside other momentum indicators. The value is not any single reading but how RSI aligns with the broader signal picture. An RSI divergence that coincides with declining volume and a MACD histogram rollover is a materially different signal than an RSI divergence on its own.

The Bottom Line

RSI is a momentum gauge, not a trading system. Stop using it as a standalone buy/sell trigger. Start using it to ask better questions: Is momentum confirming the trend? Is there a divergence forming? Has a failure swing completed?

The difference between retail traders who lose money with RSI and professionals who use it effectively is not the indicator settings or the timeframe. It is whether they understand what the number actually represents.