Academy
Momentum

RSI: Reading Momentum, Not Predicting the Future

The Relative Strength Index measures how fast and how far price has moved — not where it's going next.

6 min read

RSI (Relative Strength Index) is a momentum oscillator that compares the size of recent gains to the size of recent losses, typically over a 14-period lookback. It's plotted on a 0-100 scale. Above 70 is conventionally called 'overbought.' Below 30 is 'oversold.' Those labels are misleading if you take them literally — they don't mean price has to reverse, only that momentum has moved quickly in one direction.

What it's actually measuring

RSI is a ratio: average gain over average loss during the lookback window, rescaled to 0-100. A reading near 70 means recent up-candles have dominated recent down-candles — it says nothing about tomorrow. In a strong trend, RSI can sit above 70 or below 30 for long stretches while price keeps grinding in the same direction. Treating 'overbought' as an automatic sell signal is the single most common way traders misuse this indicator.

Three ways ChartWyz uses it

  • RSI Oversold Bounce / Overbought Reject — RSI crosses back through 30 or 70, not just touches it. The turn matters more than the level.
  • RSI Midline Bull/Bear — a cross through 50 is a cleaner momentum-shift signal than the extremes, and fires far more often in ranging markets.
  • HARSI Bullish/Bearish Divergence — RSI calculated on Heikin Ashi closes (see the Heikin Ashi lesson), watched for divergence against price rather than absolute level.

The honest limitation

RSI lags price by definition — it's built from the last 14 candles, so it confirms a move more often than it predicts one. It also generates false signals constantly in choppy, low-volatility conditions, which is exactly why ChartWyz backtests every RSI-based setup per coin and per timeframe instead of treating '30 and 70' as a universal rule. A setup with a 55% win rate on a 20-sample size is a coin flip with extra steps — check the sample size before you trust the percentage.