The Relative Strength Index is among the most popular indicators in technical analysis, and also among the most misunderstood. It measures the speed and magnitude of recent price changes, comparing the size of gains to the size of losses over a chosen period, and expresses the result on a scale from zero to one hundred. The familiar convention holds that readings above seventy mark an overbought condition and readings below thirty mark an oversold one. From this convention springs a tempting and dangerous conclusion: that an overbought reading is a signal to sell and an oversold reading a signal to buy. This interpretation is the source of more losses than almost any other idea in the field.
The flaw lies in the words themselves. To call a market overbought implies that it has risen too far and must therefore fall, but the indicator makes no such claim. A high RSI reading states only that recent gains have been large and rapid relative to recent losses. In a strong, sustained uptrend that is exactly what one would expect to see, and the indicator can remain elevated above seventy for weeks while price continues to climb. An investor who sells on the first overbought reading is betting against the very strength the indicator is reporting, and a powerful trend will punish that bet repeatedly. The same logic applies in reverse during a determined decline, where RSI can sit in oversold territory far longer than any premature buyer can endure.
What the conventional reading misses is that overbought and oversold are descriptions of momentum, not predictions of reversal. They tell you the market is moving forcefully; they do not tell you it is about to stop. The mistake is to confuse an extreme reading with an exhausted one. Strength and exhaustion can look identical at a single glance, and the difference between them only emerges over time and in context. In a range-bound market, where price genuinely oscillates around a stable centre, extreme RSI readings do tend to mark turning points, because the conditions for reversion actually exist. In a trending market they mark continuation. The indicator behaves the same way in both cases; the meaning changes entirely.
The more disciplined applications of RSI abandon the simple threshold rule altogether. Divergence between RSI and price — where price makes a new high but the indicator does not — carries genuine information, because it exposes a trend running on diminishing momentum. The behaviour of the indicator around its centre line offers a cleaner read on the prevailing balance of force than its excursions to the extremes. So-called failure swings, in which the indicator fails to confirm a new price extreme and then breaks its own recent low or high, describe a shift in momentum that the raw threshold ignores. None of these uses treats seventy or thirty as a command.
The broader lesson is one that recurs throughout technical analysis. An indicator does not become a strategy simply because it produces a number with a memorable boundary. RSI is a precise and useful measure of momentum, and a treacherous generator of mechanical buy and sell signals. Used to gauge the character and conviction of a move, and always interpreted in light of whether the market is trending or ranging, it earns its place. Used as a reflex that fades every extreme, it becomes a reliable way to stand in front of trends that have no intention of yielding.