RSI

Open almost any charting app, add the RSI, and you will see a wiggling line in a little panel below the price, with two levels marked — usually 70 near the top and 30 near the bottom. Cross above 70 and the app may shade it red and whisper overbought. Drop below 30 and it turns green and murmurs oversold.

The beginner reads this the obvious way. Overbought must mean too high, about to fall. Oversold must mean too low, about to bounce. So the plan writes itself: sell at 70, buy at 30, repeat, get rich. It is clean, it is intuitive, and in a strong trend it is a quiet disaster.

This module is about one indicator, the RSI, and one hard lesson that comes with it: in a genuine trend, the RSI can stay "overbought" for weeks while the price marches higher the entire time — and the trader who shorted the first 70 reading is now underwater and adding. Learn what the number really measures, and the trap stops working on you.

Price alone tells you where a stock is. It does not, on its own, tell you how hard it is being pushed to get there. Two stocks can both rise 5% in a week — one grinding up in tiny steady steps, the other lurching up in one violent burst and then stalling. The path matters, and a plain price line blurs it.

An oscillator is any indicator that converts price into a number that swings back and forth inside a fixed range — here, 0 to 100. Because it is bounded, it can never run off the top of the chart the way price can, which makes "high" and "low" mean something comparable across time and across stocks. Think of it as a speedometer bolted underneath the price: it does not tell you the direction you are heading, it tells you how fast you are travelling.

The RSI — the Relative Strength Index, built by J. Welles Wilder in 1978 — is the most famous of these. Its recipe, in plain words: over a lookback window (14 periods by default), it compares the average size of the up-moves to the average size of the down-moves, and squeezes that ratio onto a 0–100 scale. When up-moves have dominated, the line rides high. When down-moves have dominated, it sinks low. That is the whole idea. Everything people project onto it — tops, bottoms, reversals — is interpretation laid on top of a simple measurement of recent up-versus-down pressure.

Two things follow immediately, and holding both in mind is the difference between using the RSI and being used by it. First, because it is built only from past prices, the RSI is not a crystal ball; it is a tidy summary of what already happened. It is called a leading indicator because momentum sometimes fades before price does, but the data feeding it is entirely historical. Second — and this is the part beginners skip — a high reading is a statement about strength, and strength is exactly what a healthy trend has plenty of.

Picture the panel. The RSI line lives between 0 and 100. Two horizontal levels are drawn across it by convention:

• 70 — the overbought line. Above it, recent up-moves have strongly outweighed down-moves.
• 30 — the oversold line. Below it, down-moves have strongly outweighed up-moves.

The midline at 50 is the balance point: above 50, up-pressure has the upper hand over the window; below, down-pressure does. Many trend readers care more about which side of 50 the line sits on than about the 70/30 extremes at all.

Now the crucial correction to the beginner's instinct. "Overbought" does not mean "too high to go higher." It means momentum is strong right now. In a powerful uptrend, strong momentum is not a bug to be sold — it is the trend doing exactly what trends do. The RSI can pin itself above 70 and stay there, session after session, for as long as the buyers keep pressing. The oscillator has no ceiling of time; it only has a ceiling of 100.

The mirror image holds at the bottom. In a hard downtrend the RSI can slump below 30 and stay there while the stock keeps falling. "Oversold" in a downtrend is not a floor; it is a description of how heavy the selling has been, and heavy selling is what downtrends are made of.

So when is the 70/30 read useful? Mostly when the stock is not trending — when it is stuck in a sideways range, oscillating between a floor and a ceiling. There, RSI 70 tends to line up with the top of the range and RSI 30 with the bottom, and fading the extremes can work because the stock keeps coming back to the middle. The single most important habit with the RSI is therefore to look at the price first and ask: is this a trend or a range? The same reading demands the opposite response in each.

Walk one composite chart. illustrative

Below is a stock in a clean advance. Watch what happens if you obey the beginner's rule and short the first time the RSI would have crossed 70 — roughly the third bar of the run.

At the down arrow, the RSI first pokes above 70. The rule says short. But the buyers are not finished — they are barely warmed up. Over the following sessions the price adds roughly another 30% while the RSI simply refuses to come back below 70 for any length of time. Every day, the beginner's screen keeps flashing "overbought," and every day the trend ignores it. The short is stopped out; the "signal" was never a signal, it was a symptom of exactly the strength that was carrying the stock.

Now flip the intuition to read the same chart correctly. The persistent high RSI is confirmation, not warning: momentum agrees with price, both are pointed up, and there is no argument between them. The reading a trend-follower actually waits for is different — not the first touch of 70, but the moment the RSI finally fails to make a new high while price does, or slips back below 50 and holds there. That is momentum genuinely rolling over, and it is a far rarer, far more meaningful event than the daily "overbought" tint. This is the difference between reading the oscillator as a stop-sign and reading it as a speedometer: one screams at every fast moment, the other only matters when the speed truly changes. A trend in motion tends to keep going until something real stops it.

The RSI is a compression of the past into one tidy number, and everything it throws away on the way is something it cannot tell you.

It cannot tell you why. A 30 reading looks identical whether it came from ordinary profit-taking or from a fraud being uncovered. The oscillator sees only the size of the down-moves, never their cause — and the cause is what decides whether the fall is a dip to buy or a cliff to avoid. Momentum has no idea what the company is; it only knows how fast the price moved. The chart maps the crowd's behaviour, not the business underneath it.

It cannot tell you when the trend will actually turn. The whole "overbought stays overbought" problem is really the RSI admitting that strength and exhaustion look the same in the moment. A high reading is consistent both with a trend that runs for months more and with one about to top out. The oscillator cannot separate them, and anyone who tells you it can is selling certainty that does not exist. Every oscillator reading shifts the odds a little; none of them promises an outcome.

And it will not save you from a backtest that "always worked." It is trivially easy to find, in hindsight, the exact RSI setting and the exact 70/30 pair that would have nailed the turns on one chart. That curve-fitted rule then falls apart the moment it meets a new chart, because it was tuned to noise, not to a real edge. A rule optimised to fit the past is fitted to its accidents, and its accidents do not repeat. The honest use of the RSI is as one input among several, read in context — never as a machine you can switch on and obey.

The same handful of errors catch beginner after beginner with this one indicator.

1. Shorting every 70, buying every 30. The headline mistake. In a trend, this systematically bets against the strongest, most persistent force on the chart. The extremes are useful in ranges and treacherous in trends, and telling the two apart is a job for the price, not the oscillator.

2. Reading the label instead of the context. "Overbought" and "oversold" are just colours the app paints above 70 and below 30. They are descriptions of momentum, not instructions. The words do half the damage by themselves.

3. Treating divergence as a trigger. Divergence — price making a new high while the RSI makes a lower high, or the reverse at a low — is a real and useful caution that the push behind a move is fading. But strong trends throw off divergence repeatedly before they turn, and many divergences simply dissolve. It earns a tighter stop, not an immediate reversal trade.

4. Fiddling the settings until it "works." Changing the lookback from 14 to 9 to 21, or the levels from 70/30 to 80/20, until the indicator lines up with the past chart. This feels like refinement and is actually curve-fitting.

5. Forgetting the RSI knows nothing about the company. The oscillator is a shape made of past prices. It cannot see earnings, cash, debt, or the reason the price is moving — and those are what decide whether a trend deserves to continue.