What is Kama?
Kama, short for Kaufman’s Adaptive Moving Average, is a type of moving average that changes its smoothing level over time. Unlike a standard moving average that uses a fixed amount of smoothing, Kama aims to behave differently depending on whether recent price movement looks more “efficient” (strong, directional) or more “inefficient” (choppy, noisy).
In practice, Kama produces a line that is recalculated as new price data arrives. Traders and analysts often use this line to study trend and momentum, and to compare its response speed across different market conditions.
How Kama works
Kama is built around two ideas: an efficiency measure and an adaptive smoothing factor.
First, an efficiency ratio is estimated from recent price movement. Conceptually, it compares how much price has moved in a directionally meaningful way to how much total movement occurred over the same period. If price travels from start to end more directly relative to the amount of back-and-forth movement, the efficiency ratio tends to be higher. If price meanders and retraces frequently, the efficiency ratio tends to be lower.
Second, that efficiency ratio is converted into an adaptive smoothing constant. The smoothing constant controls how strongly the newest price influences the current Kama value.
- When the efficiency ratio is higher, Kama generally uses a larger effective smoothing constant, making the average react faster.
- When the efficiency ratio is lower, Kama generally uses a smaller effective smoothing constant, making the average smoother and less reactive.
Third, Kama is updated recursively. Each new Kama value is computed from:
- the prior Kama value (the existing smoothed baseline), and
- a fraction of the newest price change (scaled by the adaptive smoothing constant).
Two common inputs you will see in implementations are:
- a lookback period used to measure recent efficiency, and
- parameters that bound or shape the adaptive smoothing behavior (for example, values that represent a faster and a slower responsiveness).
Because different charting platforms and libraries may name these parameters differently, it helps to verify the exact formula and default settings in the specific implementation you use.
Mechanics and interpretation in a Moving Averages context
Kama belongs to the broader family of moving averages, so its output is still an averaged representation of price over time. The difference is that the “amount of averaging” is not constant.
That has a few mechanical consequences:
- Responsiveness varies with recent market character rather than being fixed.
- The distance between price and Kama can expand or compress depending on whether the efficiency ratio is currently higher or lower.
- During choppy conditions, Kama may aim to reduce whipsaw by smoothing more.
However, Kama does not guarantee correct directionality. A moving average line summarizes past data, so it can lag during sudden changes and still produce false indications when regimes shift.
Relevant limitations and risks
Kama’s adaptive nature is designed to address noise, but it also introduces practical uncertainties.
1) Parameter sensitivity Kama depends on lookback and responsiveness settings. Different values can change how quickly Kama adjusts. Without careful comparison, you may see different outcomes from what looks like the “same” indicator.
2) Market regime changes The concept of efficiency implicitly assumes that recent conditions are representative enough for the current time window. When the market transitions from trending to ranging (or vice versa), the efficiency measure may lag the new regime.
3) Overfitting risk in usage If someone tunes Kama parameters to a particular historical period, results may reflect overfitting rather than stable behavior. Even non-trading uses (like selecting parameter settings for a dashboard) can still become biased toward one timeframe or dataset.
4) Data and implementation differences Kama is sensitive to input data granularity (for example, candle size), the “price” used (close vs. typical price), and the exact formula used by the software. These differences can meaningfully alter the resulting curve.
5) No certainty from an average As with any moving average, Kama is derived from past price. It cannot observe future conditions. That means you should expect periods where it is less responsive or more misleading, especially around abrupt structural changes.
What you can verify independently
Because the goal is to understand Kama without relying on outcomes claims, focus on verifiable checks:
- Confirm the exact definition used by your platform (formula, price input, and parameter defaults).
- Compare Kama to a standard fixed moving average over the same data to see how its responsiveness changes.
- Test behavior across multiple market environments (trending, ranging, and transitions) using the same methodology.
- Check stability across different timeframes to see how sensitive results appear.
If you want, you can also use the linked internal guide topics to deepen specific aspects (such as advanced considerations, backtesting, or differentiating Kama from related concepts).