RSI Basics: What Overbought and Oversold Really Mean

When you start learning chart analysis, the very first indicator you meet is RSI (Relative Strength Index). Welles Wilder devised it in 1978, and it's built into nearly every charting app. The line "below 30 is oversold" floats around like a trading proverb. Yet surprisingly few people know exactly what that number is counting.

What RSI measures

RSI takes the ratio of up-day strength to down-day strength over a recent period and maps it onto a 0–100 scale. The default setting is 14 candles. Put into words, the calculation goes like this.

1. Over the last 14 candles, find the average of the up moves (AU) and the average of the down moves (AD) versus the prior candle.
2. Compute the relative strength RS = AU ÷ AD.
3. Convert with RSI = 100 − 100 ÷ (1 + RS).

When up-strength and down-strength are equal (RS=1), RSI is 50. If all 14 candles are up, it pins near 100. All down, near 0. So RSI isn't a device for predicting the future. It's a lagging statistic that summarizes the bias of recent price action.

The 70/30 rule and its misunderstanding

The textbook reading is simple. RSI at 70 or above is overbought (rose too much too fast). 30 or below is oversold (fell too much). In a ranging market this contrarian frame works well enough. RSI keeps crossing 70 at the top of the range and breaking 30 at the bottom, so the pattern repeats.

The problem is a trending market. In a strong uptrend, RSI can sit above 70 for weeks while price keeps climbing. Bitcoin's weekly RSI in the 2020–2021 bull run floated in the overbought zone for a long stretch. On the other side, hands that caught a crash thinking "RSI is at 30, it'll bounce" often got to watch an even deeper bottom. That's why experienced users sometimes flip the reading. They treat overbought/oversold not as a "reversal signal" but as "evidence the trend is strong." The same number allowing opposite readings is the built-in limit of indicators.

Divergence, what RSI users actually watch

One use has earned more trust than plain 70/30: divergence. If price sets a new high but RSI's peak comes in lower than before (bearish divergence), it reads as a hint that the rally's internal momentum is fading. Flip it: if price sets a new low but RSI's trough is higher (bullish divergence), you suspect the selling pressure is running out. Divergence isn't a confirmed signal either. It's a probabilistic warning light. The convention is to weight it more on larger timeframes (daily, weekly).

What to remember in practice

• The picture changes completely by timeframe. A 15-minute RSI oversold and a weekly RSI oversold carry different weight.
• RSI trading on its own has only a faint statistical edge. It's usually overlaid with trend judgment (moving averages and such) and market sentiment (Fear & Greed Index, funding rate).
• The setting of 14 is a convention, not a magic number. Shorten the period and it gets more sensitive. Lengthen it and it gets duller. Neither is "more correct."

For the record, the octopus on this site doesn't use RSI. The "recent price bias" that RSI summarizes overlaps in role with its 24-hour momentum indicator. The indicators and formulas the octopus actually reads are in the full algorithm reveal. RSI is a fine beginner's lens for understanding the market. But remember: no matter how good the lens, the future never comes out developed on the film.