The RSI Indicator, Without the Myths

The RSI indicator compresses recent price action into a single number from 0 to 100 — a fast read on momentum that retail traders routinely mistake for a buy-and-sell button. This page covers what RSI(14) actually measures, why overbought and oversold mislead more people than they help, how divergence gives a cleaner read, and why the RSI indicator on its own is not a trade signal. Research and education only — not financial advice.

The RSI indicator — Relative Strength Index — is a momentum oscillator that squeezes recent price action into one number between 0 and 100. J. Welles Wilder published it in 1978, and RSI(14), a fourteen-period lookback, is still the default on nearly every charting platform. It is genuinely useful for reading the speed and persistence of a move. It is also among the most misread tools in retail trading, because the two lines everyone draws on it — 70 and 30 — get treated as automatic sell and buy commands. They are not. What follows is what RSI actually measures, why "overbought" and "oversold" are the wrong mental model, how divergence gives a more honest read, and why the RSI indicator alone is not a signal.

What RSI(14) actually measures

RSI is a ratio of average gains to average losses over a lookback window. The formula is RSI = 100 − 100 ÷ (1 + RS), where RS is the average up-close divided by the average down-close across the period — 14 bars by default. Because it is bounded between 0 and 100, with 50 as the neutral midline, it is easy to read at a glance. That readability is exactly what makes it easy to over-trust.

Two consequences fall straight out of the math. First, RSI is relative to an instrument's own recent behavior — a reading of 65 on a sleepy utility is not the same event as 65 on a biotech that swings 10% a day. Second, the lookback length changes the tool entirely: RSI(14) on a daily chart, RSI(14) on a 5-minute chart, and RSI(7) on that same daily chart are three different indicators wearing the same name. Shortening the period makes it twitchier and drives it to the extremes more often; lengthening it smooths everything toward 50. There is no universally "correct" setting — only a setting matched to a timeframe and a purpose.

The overbought / oversold myth

Convention labels RSI above 70 "overbought" and below 30 "oversold," and the folk interpretation is that overbought means sell, oversold means buy. This is where a lot of small accounts go to die. Overbought is not a ceiling — it is a description of strength. In a powerful uptrend, RSI can pin above 70 for weeks while price keeps grinding higher; mechanically shorting every 70 print means fading the strongest tape in the market. Symmetrically, a stock in freefall can sit under 30 for a long time before it finds a floor, and "oversold" catches every falling knife on the way down.

RSI reading	Naive reading	More accurate reading
Above 70	"Overbought — sell"	Recent move has been fast and one-sided; strength, not a ceiling
Below 30	"Oversold — buy"	Selling has been fast and one-sided; weakness, not a floor
Around 50	"Neutral"	No momentum edge either way — often the least tradable state

The honest translation of a high RSI is simply: the recent move has been rapid and mostly in one direction. Whether that resolves as a reversal or a continuation depends on context the oscillator cannot see — trend, catalyst, and participation. In a strong trend, professionals often read a high RSI as confirmation of momentum, not a reason to fade it.

Divergence: the more useful read

Divergence is where RSI earns its keep, because it surfaces something price alone doesn't say out loud. It occurs when price and the oscillator disagree. Bearish divergence: price prints a higher high, but RSI prints a lower high — the new high was made on weaker momentum, a hint the move is tiring. Bullish divergence: price makes a lower low while RSI makes a higher low — selling pressure is fading even as price slips. Divergence describes the quality of a move, not just its direction.

The critical caveat: divergence is a condition, not a timing tool. Momentum can fade and keep fading while price extends for a long time — a market can stay divergent far longer than an impatient trader can stay solvent. That is precisely why divergence belongs on a watch-for list, waiting for a separate confirmation, rather than being traded the instant it appears.

Why RSI alone is not a signal

Every objection above shares one root cause: RSI is a lagging transform of price. It contains no information price didn't already carry — it just repackages the last 14 bars into a smoother shape. An indicator built entirely from past price cannot, by construction, know the future. Layering a second price-derived indicator on top — say MACD — doesn't add independent evidence; it mostly adds correlated noise. What actually changes the odds is context RSI can't see: a catalyst, a level, and real participation. Relative volume answers a question RSI never can — is anyone actually here for this move, or is it drifting on nothing?

There is data behind the skepticism. In a published hypothetical backtest, a raw mechanical scanner traded blind — the kind of "the number crossed my line, so I click" logic RSI invites — produced a 46.6% simulated win rate and a 0.82 profit factor, i.e. a negative expectancy per trade. A single indicator crossing a threshold is a starting question, never a finished answer; you can read the full breakdown at the public record.

The trap in one line: RSI hitting 70 or 30 tells you a move was fast — it does not tell you the move is over. Treat the oscillator as a description of momentum, not as an instruction.

Where RSI actually fits

Used well, RSI is a filter and a context gauge, not a trigger. It helps you characterize a setup — is momentum accelerating, stalling, or diverging? — before a separate, price-based event tells you the idea is live. On our desk, RSI is one input into a card, never the whole card. A complete idea specifies a trigger, take-profit targets, a stop, and a time-stop, posted before the move to a timestamped paper record where the losers stay visible — you can see how that discipline reads on live setups in the signals feed. The oscillator informs the read; it does not place the trade. That distinction — indicator as context versus indicator as command — is the difference between using RSI and being used by it.

Common questions

Is RSI above 70 a sell signal?

No. RSI above 70 is conventionally called "overbought," but that only describes a fast, one-sided move — it is not a ceiling. In a strong uptrend RSI can hold above 70 for weeks while price keeps rising, so mechanically shorting every 70 reading means fading the strongest tape in the market. It is a description of momentum strength, not an instruction to sell.

What is the best RSI setting?

There is no universal best. RSI(14) is Wilder's default and the most common, but the right period depends on your timeframe and goal. Shorter lookbacks (e.g. RSI 7) react faster and hit the extremes more often, which suits short-term trading but produces more false extremes; longer lookbacks smooth toward the 50 midline. RSI(14) on a daily chart and RSI(14) on a 5-minute chart are effectively different tools.

What is RSI divergence?

Divergence is when price and RSI disagree. Bearish divergence is a higher high in price paired with a lower high in RSI — a new high on weaker momentum. Bullish divergence is a lower low in price with a higher low in RSI — a new low on fading selling. It describes the quality of a move, but it is a condition, not a timing tool: markets can stay divergent for a long time, so it needs separate confirmation.

Can you trade profitably with RSI alone?

RSI is a lagging transform of past price, so by construction it holds no information price didn't already contain — which is why an indicator crossing a threshold is a starting question, not a finished signal. This page treats RSI as a context filter used alongside a catalyst, a level, relative volume, and a defined trigger, stop, and target. It is an educational framework, not a promise about any outcome, and not advice to buy or sell anything.

See the process with your own eyes. The desk posts trigger-based cards to a public, timestamped record — losses included — and published the backtest where its own raw scanner loses. The scoreboard is free to watch. Join the floor →

Free to join · paid floors optional · research and education only

Last updated 2026-07-11 · ClaudeQuantAlgo Research Desk · research and education only.

Disclosures. ClaudeQuantAlgo is a research and education community. Nothing on this page constitutes financial, investment, legal, or tax advice, or a recommendation to buy or sell any security or derivative. No profit promises are made, ever — trading stocks, options, and forex involves substantial risk of loss; options positions can lose 100% of their value. The public scoreboard reflects a model ("paper") desk — no real money. Past performance — real, paper, or simulated — never guarantees future results. Hypothetical and simulated results have inherent limitations and no representation is made that any account will or is likely to achieve similar profits or losses. ClaudeQuantAlgo is not a registered investment adviser or broker-dealer. You are solely responsible for your own trading decisions. Never risk money you cannot afford to lose.