What RSI actually measures
RSI was introduced by J. Welles Wilder Jr. in 1978. It compares the size of recent gains to the size of recent losses over a lookback window — most commonly 14 periods — and expresses the result as a value from 0 to 100. Loosely, a high reading means recent up-moves have dominated recent down-moves; a low reading means the opposite.
It is important to be precise about what this is: RSI is a smoothed ratio of past price changes. It contains no information about news, order flow, liquidity, or the future. Every RSI value is fully determined by prices that have already happened, so the indicator by construction lags the market it describes.
How traders commonly read it
The best-known convention is the 70/30 rule: readings above 70 are labelled 'overbought' and readings below 30 'oversold.' Some traders shift these thresholds (e.g., 80/20) in strongly trending markets. Traders also watch for RSI crossing the 50 midline, and for 'divergence' — where price makes a new high or low but RSI does not, which some interpret as weakening momentum.
These are conventions, not laws. 'Overbought' does not mean 'about to fall' and 'oversold' does not mean 'about to rise.' The labels describe momentum that has been strong or weak recently; they say nothing reliable about what happens next. Different traders use different settings and thresholds, which means the same chart can produce contradictory readings depending on who is looking.
The well-known false signals
The most famous failure mode is that RSI can stay overbought or oversold for a long time. In a powerful uptrend, RSI can pin above 70 for weeks while price keeps climbing; acting on the 'overbought' label would mean fighting the trend repeatedly. The same happens in reverse during sharp declines. This is not a rare edge case — it is a routine feature of trending markets, which Bitcoin often exhibits.
Divergence signals are similarly unreliable: RSI can diverge from price many times without any reversal following, so divergences frequently 'fail.' RSI can also whipsaw — flipping above and below thresholds in choppy, sideways conditions — generating a stream of signals that cancel out. And because RSI is calculated from a fixed lookback, a single large candle can swing the reading sharply, producing a jump that reflects the arithmetic of the formula more than any meaningful shift.
Why it is contested and subjective
RSI, like chart patterns, Fibonacci levels, and Elliott Wave, is a method of technical analysis whose predictive value is contested. There is no robust, consistent evidence that RSI thresholds reliably forecast future prices. Its output depends heavily on choices the user makes — the lookback length, the thresholds, the timeframe — and those choices are subjective. Two analysts can look at the identical price series and reach opposite conclusions.
Bitcoin adds its own difficulties: it trades 24/7 with no closing bell, and it can move violently on thin liquidity. That means RSI readings can shift fast and 'extreme' values can occur far more often than the tidy 70/30 framing suggests.
Using RSI honestly, and where data context fits
If RSI is used at all, it is best understood as one descriptive lens on recent momentum — not a signal generator and never a substitute for understanding risk. It does not tell you position size, it does not tell you when to buy or sell, and it cannot manage the possibility that you are simply wrong. Trading is high-risk, and studies of retail traders consistently find that most lose money over time.
BIKENZO is a Bitcoin data and analytics terminal, not a broker, exchange, or advisory service, and it makes no predictions. Where it can add genuine context is by showing market-data such as liquidity alongside the Bitcoin price — for example, how thin or deep the order environment is when a price move happens. That kind of context can help you interpret why a move occurred, but it is not a forecast and not a trading signal. Any decision, and any loss, remains entirely your own.