Direct answer: “How often is RSI right” in forex?
There is no universal number for how often the Relative Strength Index (RSI) is “right” in forex. RSI produces readings (for example, above or below specific levels), but “right” depends on a separate rule for what outcome counts as correct—such as whether price reverses within a set number of bars, whether RSI crossovers align with momentum turning points, or whether an overbought/oversold condition leads to a move in a particular direction.
Because different traders and researchers choose different definitions, timeframe lengths, and thresholds, any single success rate would be arbitrary unless it is measured with the exact same setup on historical data.
How RSI works in the first place (and why that affects “right” frequency)
RSI is a momentum oscillator built from recent price changes. The most common version uses a 14-period calculation and returns a value between 0 and 100. It is often interpreted through thresholds, such as:
- RSI above a chosen upper level (commonly associated with “overbought”)
- RSI below a chosen lower level (commonly associated with “oversold”)
Bollinger Bands and RSI are sometimes combined to add context: Bollinger Bands describe statistical movement around a moving average, while RSI adds momentum/relative strength. Even then, the combined idea still needs a clear, testable condition to label a “correct” event.
The key reason success frequency is not fixed: RSI can remain elevated or depressed for extended periods when price moves in a sustained trend. So an RSI signal that looks like “overbought” can stay “overbought” without causing an immediate reversal.
Example checks: how to measure “how often RSI is right” without guessing
If you want an independently verifiable estimate, you can measure it from data using explicit rules. The main choice is the scoring method:
-
Define the event Example: “RSI crosses above 30” (or below 70), using a specific RSI length and specific thresholds.
-
Define what “right” means Example: price moves in the expected direction by at least X units (or X% relative move) within N bars after the event.
-
Define the timeframe and market conditions RSI behavior differs across short timeframes versus longer ones, and volatility differs across currency pairs.
-
Repeat across enough samples A measured success rate from one month or one pair can change materially when the dataset changes.
-
Compare alternatives under the same scoring rule To understand whether RSI adds value, compare RSI-triggered events against another baseline (for instance, a simple “no-signal” expectation) using the same evaluation window.
This approach turns the question into something measurable. Without these definitions, “how often RSI is right” can only be answered qualitatively.
Limitations and uncertainty to keep in mind
- RSI success is not guaranteed: it is a rule-based indicator whose output is based on past changes, not future outcomes.
- Any “percentage correct” is conditional: changing thresholds, RSI period, evaluation horizon (N bars), or event definition changes the result.
- Overbought/oversold interpretations can be misleading during strong trends, because RSI can stay in extreme territory.
- No real-time performance claim is possible from theory alone; you only know the frequency after testing with a clear, repeatable protocol.
If your goal is “how often,” the only defensible answer is the one you measure for your exact definition of a correct outcome, your timeframe, and your dataset.