The Relative Strength Index (RSI) measures how one-sided recent price movement has been. It compares the average size of up-closes to down-closes over the last 14 candles and squeezes the result into a 0–100 scale. Readings above 70 are labelled "overbought," below 30 "oversold."
Those labels are responsible for more blown beginner accounts than almost any other piece of trading vocabulary, because they sound like instructions. They are not.
Overbought is a description, not a signal
RSI above 70 means buyers have dominated the last 14 candles. That is all. In a genuine uptrend, that is exactly what you would expect to see — strong trends live above 70 for days or weeks. Shorting a market because RSI is "overbought" is shorting strength, and strength is precisely the thing that can keep going far longer than your stop loss can survive.
The honest reframe: in a range, extreme RSI readings near known support and resistance mark areas where a swing back toward the middle is more likely. In a trend, extreme readings mark momentum, and fading them is standing in front of traffic. The indicator did not change between those two sentences — the context did. Deciding which regime you are in comes first, and that is a market structure question RSI cannot answer for you.
Divergence: the one RSI pattern worth learning properly
Divergence occurs when price makes a new extreme but RSI does not confirm it. Price prints a higher high while RSI prints a lower high: bearish divergence. Price makes a lower low while RSI makes a higher low: bullish divergence.
What divergence actually tells you is that the latest push carried less force than the one before it — the trend is doing more work for less progress. That is genuinely useful information. What it does not tell you is when anything will happen. Trends can print divergence three, four, five times before finally turning, which is why traders who short every bearish divergence in a strong uptrend get stopped out repeatedly and conclude the tool is broken.
The disciplined use: treat divergence as a warning light, not a trigger. When you see it, tighten your management of existing trend-following positions and start watching for an actual reversal signal — a break of structure, a failed high, a rejection at a key level. Enter on the structural evidence, with the divergence as supporting context. Divergence alone is a reason to pay attention, never a reason to click.
The 50 line: RSI's most underrated level
Between the famous 70 and 30 sits the level professionals quietly use more: 50. RSI holding above 50 means up-closes are consistently outweighing down-closes — momentum is bullish. Repeated rejection from the 50 area during pullbacks is characteristic of a healthy trend, and a decisive loss of 50 after a long stretch above it is often an earlier, calmer heads-up than any overbought reading.
Used this way, RSI becomes a momentum filter rather than a reversal oracle: longs preferred while RSI ranges above 50, shorts preferred below it, stand aside when it oscillates around the line. Simple, boring, and far more robust than hunting extremes.
Settings, and why you should mostly leave them alone
The default 14-period setting is fine. Shortening it to 7 or 9 makes the line more responsive and much noisier; lengthening to 21 smooths it at the cost of lag. None of these transforms a losing approach into a winning one, and cycling through settings in search of the version that would have caught the last move is the same curve-fitting trap covered in backtesting basics. Pick the default, learn its personality across a hundred trades, and spend your optimisation energy on risk management instead — it pays better.
Where RSI fits in a real trading plan
A grounded workflow looks like this. Establish trend and structure first, from the chart itself. Use RSI's position relative to 50 as a momentum filter agreeing or disagreeing with that read. Watch for divergence at extremes as an alert to manage risk. Then execute only on your actual entry criteria — structure, levels, confirmation — with risk fixed per trade using the position size calculator, because no indicator reading justifies oversizing.
Notice what is absent from that workflow: buying because RSI touched 30. An indicator can support a decision; the moment it makes your decisions, you have outsourced your edge to a formula that everyone else's platform draws identically.
One more expectation worth setting: no RSI technique changes the fact that losing trades and losing weeks are part of every honest approach. A momentum filter shifts the odds slightly in your favour over a large sample; on any given trade it guarantees nothing, and treating a "perfect" RSI reading as certainty is how sensible sizing gets abandoned at the worst moment.
RSI will not make you profitable, and neither will any other indicator — profitability comes from a plan with positive expectancy executed with discipline through inevitable losing stretches. What RSI can do is make a good process slightly better: one more honest witness to interview before you put money at risk. Keep it in that role and it earns its place on the chart.
Education only, not financial advice. Trading carries risk of loss; never trade money you cannot afford to lose.
