Why your RSI lies (and what to use instead)

The Relative Strength Index is the first indicator nearly every new trader learns. It's also the first one to lose them money. Here's what's actually going wrong — and a better mental model that takes ten minutes to internalize.

The thing every trading guru tells you about RSI

Open YouTube. Search "how to trade RSI." Watch any three videos. You'll hear the same script:

"When RSI goes above 70, the asset is overbought — that's your signal to sell. When RSI drops below 30, it's oversold — that's your signal to buy."

It sounds clean. It fits on a sticky note. It is also, in practice, one of the most expensive ideas in retail trading.

Run this trade on SPY from January 2024 through August 2024 — short every time RSI crossed above 70, long every time it dipped below 30. You'd have shorted the relentless first-half rally on RSI > 70 at least eleven times, getting stopped out on most of them. You'd have caught maybe one of the August lows. Net result: a portfolio in deep red while the index returned around 17%.

RSI didn't malfunction. The model people teach about RSI is broken.

What RSI actually measures

RSI is, mathematically, the ratio of average gains to average losses over a lookback period (default 14 bars), normalized to a 0–100 scale. That's it. Nothing about it predicts a reversal. Nothing about it tells you the asset is "overbought." It just summarizes how aggressively the price has been moving up versus down over the last fourteen bars.

In a strong uptrend, that ratio is supposed to be high. RSI will sit between 60 and 80 for weeks. It's not telling you "the move is exhausted" — it's telling you "the move is strong." Selling because the trend is strong is the opposite of what you want to do.

The popular interpretation treats RSI as a reversal signal. It's actually a momentum strength indicator. Those are very different things, and the confusion costs traders real money.

The three things RSI cannot see

1. The higher timeframe

An RSI of 75 on the 5-minute chart could mean anything. Maybe the daily chart has been in a textbook uptrend for three months. Maybe the daily just broke down and we're in a vicious bear rally. Same RSI value, completely different setups. A 5-minute indicator looking only at 5-minute data is structurally blind to what's actually driving the chart.

This is why "multi-timeframe analysis" gets repeated like a mantra — but most traders never operationalize it. They just glance at the daily chart for context, then go back to staring at the 5-minute. There's no clean way to combine what the daily is saying with what the 5-minute is saying using a single indicator.

2. Volatility regime

RSI doesn't care whether the market is in a tight consolidation or a fast-expansion breakout. But these are completely different trading environments. A momentum signal during a low-volatility chop has roughly zero predictive power. The same momentum signal during a volatility expansion is statistically meaningful.

Without measuring volatility regime separately, RSI gives the same readout in both situations. You have no idea whether to trust it.

3. Where the chart actually cares about price

Markets remember levels. Yesterday's high. The pre-market low. The 200-day moving average. A swing high from three months ago that hasn't been touched. When price reaches one of these levels, real money makes real decisions. The level matters to the math whether or not it matters to your indicator.

RSI doesn't know any of these levels exist. A buy signal from RSI five cents above a major resistance is a worse trade than the same signal in the middle of a clear range. RSI shows them as identical.

The mental model that works

Instead of treating RSI as a one-dimensional reversal signal, think of any chart as having four independent dimensions:

1. Higher-timeframe trend — is the bigger picture leaning up, down, or sideways?
2. Momentum — is there force behind the current move? (This is where RSI actually helps.)
3. Volatility regime — is the market expanding or contracting, and in which direction?
4. Key level proximity — is the chart near a price the market historically cares about?

RSI gives you exactly one of these. The other three matter just as much.

A high-quality trade is when several of these dimensions agree. The daily trend is up, momentum is rising, volatility is expanding to the upside, and we just broke a clean key level. That's a setup with asymmetric edge.

A low-quality trade is when these dimensions disagree. Momentum says go, but trend says wait. Volatility says expansion, but levels say we're rejecting overhead resistance. This is "fight zone" — and it's where most retail accounts die.

What to actually do with RSI

RSI is still useful. It's a clean way to measure momentum on the chart you're trading. The fix isn't to throw it out — it's to stop asking it questions it can't answer.

• Use RSI for what it's good at: measuring whether momentum is rising or falling, and how strong it is right now.
• Don't treat RSI extremes as reversal signals. RSI 80 in an uptrend is normal. RSI 80 against the trend is a different story — but you need to know the trend, which RSI doesn't tell you.
• Combine it with separate measurements of higher-timeframe trend, volatility, and level proximity. These are independent dimensions and they have to be measured separately.
• Wait for agreement. Most of the time these dimensions will disagree — and the right action is to do nothing. The hardest skill in trading is sitting out.

The honest summary

RSI didn't lie to you. Whoever taught you that "RSI 70 = sell" did. Once you stop using RSI as a one-button signal and start treating it as one of four dimensions that have to agree before you trade, the indicator becomes useful again. And about 80% of the time, you'll find the dimensions don't agree — which is a feature, not a bug. That's the math telling you to wait.