Direct answer
Changing settings for RSI and a moving average mostly changes sensitivity and lag. A shorter lookback window usually reacts faster to recent price changes but can also produce more short-term fluctuations. A longer window usually smooths more and reduces noise, but it tends to respond more slowly. When you combine RSI with a moving average, these trade-offs affect how quickly the indicators “agree,” how often they diverge, and how sensitive the visuals are to market regime and data conditions.
Mechanism: what the settings actually change
RSI (Relative Strength Index) is an oscillator that compares recent upward and downward price movement over a chosen lookback length (commonly expressed as “periods”). In a simplified sense, the RSI value depends on how you measure “recent” gains versus losses. Changing the RSI lookback length shifts the balance between responsiveness and stability:
- With a shorter lookback, the “recent” window is smaller, so RSI moves more when prices change sharply.
- With a longer lookback, RSI reflects a broader history, so the curve tends to be steadier.
A moving average (MA) is a line derived from past prices using a chosen averaging window. Changing the MA length changes how much older data still influences the current value:
- A shorter MA typically tracks price changes more closely, reducing lag but increasing sensitivity to noise.
- A longer MA smooths more heavily, increasing lag but dampening short-term swings.
When you use both together, the RSI’s sensitivity and the MA’s responsiveness become coupled. For example, a faster MA can “follow” price changes sooner, while a slower RSI may still reflect prior conditions—so the two may appear to disagree more often even if both calculations are internally consistent.
Evidence or example (with explicit assumptions)
Assume you are using a fixed price series and only changing one parameter at a time.
- Suppose there is a sudden one-week move that creates larger-than-usual gains followed by a period of stabilization. If you shorten the RSI lookback, RSI will typically reach more extreme levels sooner, because it weights a smaller set of gains/losses more heavily. With a longer lookback, the same move spreads across a larger history window, so RSI may rise less sharply and then fall more gradually.
- Now assume the trend reverses after the stabilization. If you use a short MA, it often turns earlier because fewer past prices influence the average. With a longer MA, the average can stay near the previous direction longer, since it still includes more older observations.
These are not guarantees—different price paths can produce different shapes—but they illustrate the general sensitivity-versus-lag effect that parameter changes introduce. Historical behavior also depends on the exact price series used (for example, whether you average closes consistently) and on consistent calculation rules across the periods.
Limitations and risks: why settings can mislead
- Parameter overfitting: Trying many combinations of RSI and MA settings to match a past period can create a model that looks good historically but does not generalize.
- Regime changes: Market behavior can shift (volatility, trendiness, liquidity). Settings that work well in a trending regime may react too quickly—or too slowly—when conditions change.
- Indicator mismatch: Using two indicators does not remove the underlying issue that both are derived from price history. They can diverge because their lookbacks and smoothing respond differently to the same event.
- Practical frictions: Any real-world outcome also depends on costs, execution quality, and jurisdictional rules. Even if indicator behavior is consistent, the net result may differ.
Verification and next question
To independently verify how settings change RSI and moving averages, keep the data and calculation method constant and vary only one setting at a time (for example, RSI lookback while MA settings stay fixed, or MA length while RSI stays fixed). Compare how quickly the lines respond after known price events, and document where the indicators become noisy versus where they become slow to react. Then repeat across multiple historical periods to check whether the behavior is consistent.