"RSI is oversold — time to buy" may be the most expensive sentence in retail trading. The Relative Strength Index is a genuinely useful tool, but it answers a much narrower question than most people think it does, and the difference between markets where it works and markets where it lies is not subtle. It is the difference between a range and a trend.
The Relative Strength Index (RSI) is a momentum oscillator that compares the size of recent gains to the size of recent losses over a lookback window — 14 periods by default — and compresses the answer into a scale from 0 to 100. High readings mean recent candles have been dominated by gains; low readings mean losses have dominated.
By convention, readings above 70 are called overbought and below 30 oversold. Those labels are the problem. RSI does not measure whether an asset is "too expensive" or "too cheap" — no indicator can. It measures how one-sided recent price movement has been. Whether one-sided movement is about to reverse or about to continue depends entirely on context the indicator cannot see.
The classic advice — buy near 30, sell near 70 — is a mean-reversion play. It assumes price oscillates around a stable centre, so an extreme reading implies a stretched rubber band that will snap back.
In a genuine ranging market — price bouncing between well-defined support and resistance with no directional drift — that assumption often holds. Extremes in RSI tend to line up with the edges of the range, and fading them (trading against the recent move) has a logic: the market keeps returning to the middle. RSI earned its reputation in exactly these conditions.
Now put the same rule inside a strong downtrend. Price falls from $100 to $80; RSI drops to 28 — oversold. You buy. Price bounces weakly to $82, then slides to $70. RSI reads 25 — even more oversold. You add. Price grinds to $58. At no point did the indicator malfunction: losses genuinely kept dominating gains. That is what a downtrend is.
In trending markets, RSI can sit below 30 — or above 70 in an uptrend — for weeks. The extreme reading stops being a reversal hint and becomes a strength gauge: persistent overbought is what powerful uptrends look like from inside. Traders who fade every extreme in a trend are not buying value; they are standing in front of momentum, repeatedly, with the indicator cheering them on.
This is why the same number needs two readings:
| RSI reading | In a ranging market | In a trending market |
|---|---|---|
| Below 30 ("oversold") | Price near range support; a bounce is plausible | Downtrend is strong; continuation is the default expectation |
| Above 70 ("overbought") | Price near range resistance; a fade is plausible | Uptrend is strong; continuation is the default expectation |
| Holding 40–60 | Mid-range chop; little information | A pullback pausing near trend equilibrium — often where trends resume |
Classify the market first, then read the indicator. The regime question — is this market trending, ranging, or just volatile? — decides which interpretation applies, which is why we cover it separately in the guide to market regimes. An RSI signal without a regime read is a coin flip wearing a number.
In trends, use RSI with the trend, not against it. In healthy uptrends, RSI rarely reaches 30; pullbacks tend to bottom in the 40–50 zone. A dip into that zone at a sensible price level — old resistance retested as support, say — is a far better use of the indicator than hunting reversals. How those levels form and hold is covered in support and resistance.
Never let RSI act alone. One derived-from-price number should be one vote among several independent ones — trend alignment, location, volume, trade quality. That voting discipline is the entire subject of confluence, and it exists in large part because of how often oscillators mislead in the wrong regime.
It is worth seeing how a rules-based system defuses this trap, because the mechanics are instructive. In SRA Quant's published scoring methodology, RSI belongs to a momentum category worth at most 20 points of a much larger confluence score — and momentum points are only awarded when the reading agrees with the proposed trade direction. An oversold reading contributes nothing toward a long idea if the higher-timeframe trend contradicts it; the trend category is zeroed outright when the 4-hour trend disagrees with the trade. The engine also classifies the market regime on every analysis, precisely so that an oversold print in a falling market is not dressed up as a reversal signal — the reasoning the how-it-works page describes in plainer terms.
You do not need software to borrow the principle: decide what kind of market you are in before you decide what an indicator is telling you, and let no single reading — however extreme — carry the decision alone.
SRA Quant applies these principles automatically. Momentum readings are weighed against trend, regime, levels and volume before they can influence any alert — so oversold in a downtrend stays what it is: a warning, not an invitation.
SRA Quant provides market analysis and educational content only. Nothing on this page or the platform constitutes financial advice, and past or backtested performance does not guarantee future outcomes. Trading involves substantial risk of loss.