Kama in simple terms
KAMA stands for Kaufman’s Adaptive Moving Average. It is a type of moving average designed to change how “smooth” it is over time. In practice, that means KAMA can behave more like a slower average during less-trending conditions, and more like a faster average when price changes become more persistent.
Because KAMA adapts its smoothing, it is often grouped with moving-average indicators. However, it is not just a standard moving average with a fixed lookback. The key idea is that KAMA tries to link its responsiveness to a mathematical measure of how efficiently price is moving.
How Kama works: inputs and mechanics
KAMA is computed from price history using several steps and assumptions:
- Choose a price series (for example, a closing price) and a base smoothing setup. Different implementations may allow different price inputs.
- Compute an efficiency ratio (often based on how far price has moved compared with the total amount of movement over a lookback window).
- Convert that efficiency into a smoothing constant. Higher efficiency typically produces a larger smoothing factor, making KAMA react more quickly.
- Iterate the moving average: KAMA at the next time step updates from the prior KAMA value plus a fraction of the difference between current price and prior KAMA.
A simple way to think about the mechanics is: the algorithm decides the weight it gives to the newest price, and that weight changes over time based on the efficiency ratio.
Evidence and examples you can verify
A useful check is to run a small, fully documented calculation on a short price sequence.
For example, assume you have a sequence of prices over a fixed lookback window and you compute:
- the net change (distance from the first price to the last price),
- the sum of absolute changes (how much price moved in total within the window),
- an efficiency ratio derived from those two quantities,
- then a smoothing constant derived from the efficiency ratio,
- and finally the recursive update for KAMA.
If two charting platforms claim they compute KAMA, you can independently verify whether they match by comparing:
- which price field they use (close vs. another price),
- how they define the efficiency ratio,
- how they transform efficiency into the smoothing constant,
- and how they handle the early bars where the lookback window is incomplete.
This kind of verification matters because KAMA can appear to differ across providers even when the name is the same.
Limitations and risks
KAMA’s adaptability is also its main limitation. The same adaptive behavior can help in some conditions and hurt in others.
Material failure modes include:
- Efficiency ratio sensitivity: If your market produces noisy swings, the efficiency measure may overestimate or underestimate persistence, changing KAMA’s responsiveness in ways you did not intend.
- Parameter dependence: Results vary with lookback length and any platform-specific implementation choices. Different parameter sets can produce noticeably different lines.
- Non-predictive behavior: A moving average line reflects past prices. It does not inherently predict future direction, and historical fit does not guarantee future performance.
There are also non-model factors that affect any evaluation: trading costs, execution quality, and data handling can change outcomes when you move from chart inspection to any real-world usage.
Verification and next questions
If you want to confirm that KAMA on your platform matches the intended concept, focus on what you can verify without relying on claims:
- Confirm the exact definition of its efficiency ratio and smoothing constant.
- Check how it handles the first computable values.
- Compare outputs across at least two tools using the same input data and parameters.
- Validate assumptions through structured testing that includes realistic costs and execution assumptions.
If you are currently comparing KAMA to other moving averages, a helpful next question is how KAMA differs from fixed-window averages and what that implies for responsiveness during regime changes.