Kaufman Adaptive Moving Average (KAMA) Explained

Direct Answer

The Kaufman Adaptive Moving Average (KAMA) is a moving average, developed by Perry Kaufman, that automatically speeds up during a strong trend and slows down during a choppy range. It does this by measuring an Efficiency Ratio, how much of the recent price movement was net directional progress versus back-and-forth noise, and using that ratio to scale its own smoothing constant between a fast setting and a slow setting.

Key Takeaways

• KAMA adjusts its own smoothing speed using the Efficiency Ratio (ER), which ranges from 0 (pure noise) to 1 (a straight-line trend).
• ER is commonly calculated over a 10-period lookback, comparing net price change to the sum of every individual price move.
• The Smoothing Constant (SC) scales between a fast 2-period EMA constant and a slow 30-period EMA constant, squared, based on ER.
• In a strong trend KAMA behaves like a fast-moving average; in a choppy range it behaves like a slow one, which is meant to reduce whipsaws.
• KAMA is still a lagging indicator built entirely from past prices, it describes recent trend efficiency, not future direction.

What Is the Kaufman Adaptive Moving Average?

KAMA was developed by Perry Kaufman as a moving average that adjusts its own smoothing speed based on market noise. A standard simple or exponential moving average uses the same fixed period no matter what the market is doing, a 20-period EMA reacts the same way whether price is trending cleanly or chopping sideways. KAMA instead measures how efficiently price has been moving over a recent lookback window and uses that measurement to decide, bar by bar, whether to behave like a fast average or a slow one.

The core input to that decision is the Efficiency Ratio (ER), sometimes called the fractal efficiency ratio. It compares the net distance price has traveled over the lookback period to the total distance price actually covered getting there (including every back-and-forth move). A market that goes straight from point A to point B has an ER near 1; a market that zigzags in place without making net progress has an ER near 0.

The Formula

Efficiency Ratio (ER) = |Close − Close n periods ago| ÷ Sum of |Close_i − Close_i−1| over the same n periods, where n is typically 10. ER ranges from 0 (pure noise/choppy) to 1 (a straight-line trend).

Smoothing Constant (SC) = [ER × (fastest SC − slowest SC) + slowest SC]². The fastest SC commonly uses a 2-period EMA constant (2 ÷ (2+1) = 0.6667) and the slowest SC commonly uses a 30-period EMA constant (2 ÷ (30+1) = 0.0645).

KAMA_today = KAMA_yesterday + SC × (Price − KAMA_yesterday).

Squaring the scaled ER inside the SC formula pulls the constant toward the slow end for most middle-of-the-road ER values, so KAMA only speeds up meaningfully once a trend is fairly clean. In a strong trend (ER near 1), SC approaches the fast constant and KAMA moves quickly, closely tracking price like a short-period average. In a choppy range (ER near 0), SC approaches the slow constant and KAMA moves slowly, behaving like a long-period average, which is the mechanism meant to reduce whipsaws during sideways conditions.

Worked Example

Hypothetical example, for education only.

Suppose over the last 10 periods a stock's close moved from $98.00 to $104.00, and the sum of every individual absolute period-to-period price change over those 10 periods was $10.00 (reflecting some back-and-forth along the way, not a perfectly straight climb).

ER = |104.00 − 98.00| ÷ 10.00 = 6.00 ÷ 10.00 = 0.60

Using the commonly cited fastest SC of 0.6667 and slowest SC of 0.0645:

SC = [0.60 × (0.6667 − 0.0645) + 0.0645]² = [0.60 × 0.6022 + 0.0645]² = [0.3613 + 0.0645]² = 0.4258² ≈ 0.1813

If yesterday's KAMA value was $101.50 and today's price is $104.00:

KAMA_today = 101.50 + 0.1813 × (104.00 − 101.50) = 101.50 + 0.1813 × 2.50 = 101.50 + 0.45 ≈ $101.95

With an ER of 0.60, a moderately efficient move, not a clean straight-line trend, the resulting smoothing constant sits well below the fastest possible setting, so KAMA moves only part of the way toward today's price rather than tracking it closely.

How KAMA Is Commonly Used

Trend-following crossovers

Because KAMA speeds up in efficient trends, some traders use price crossing above or below KAMA, or a fast KAMA crossing a slower KAMA, similarly to how a moving-average crossover is read on a traditional EMA, treating a cross as a possible trend-direction change. As with any moving-average crossover, this confirms a move already underway rather than predicting it.

Flattening as a range filter

When ER stays low and KAMA flattens out, that flatness itself is sometimes read as a signal that the market is in a low-efficiency, choppy phase where trend-following signals from other indicators may be less reliable. A KAMA that's actively sloping, by contrast, suggests price is currently making efficient directional progress.

Dynamic support/resistance

Like other moving averages, some traders watch KAMA as a dynamic level where price tends to find support in an uptrend or resistance in a downtrend, though this behavior isn't unique to KAMA and isn't guaranteed to hold on any given touch.

Common Parameters

These are commonly cited defaults from Kaufman's original design, not universal constants, verify the exact inputs your specific charting platform uses, since some implementations expose the lookback and the fast/slow constants as adjustable settings.

Limitations

• Still a lagging indicator, every input (ER and the moving-average calculation itself) is built from past closing prices, so KAMA confirms a change in trend efficiency after it has already started.
• The ER lookback can mask short-term reversals, a market that just turned choppy after a clean 10-period trend may still show a high ER for several more bars until the noisy period fully enters the lookback window.
• Whipsaws aren't eliminated, only reduced, KAMA is designed to slow down in noisy conditions, but a sharp, brief move within an otherwise choppy range can still trigger a false signal.
• No single setting is optimal across assets or timeframes, the same lookback and fast/slow constants can behave very differently on a low-volatility large-cap stock versus a volatile small-cap or crypto asset.

Common Mistakes

• Treating a KAMA crossover as a standalone trade trigger, like other moving-average crossovers, it works best combined with broader trend context, support/resistance, or volume.
• Assuming a low ER always means "do nothing", a low reading describes recent inefficiency, not a guarantee that the range will continue.
• Comparing KAMA values across differently priced assets, like any price-based moving average, KAMA is expressed in the underlying asset's price units and isn't directly comparable between a low-priced and high-priced instrument.
• Assuming KAMA eliminates lag entirely, it reduces lag during efficient trends compared to a fixed slow average, but it does not remove lag altogether, since it is still built from historical price data.

An Average That Changes Speed Based on Recent Noise

KAMA varies its smoothing according to an efficiency measure, moving quickly when price is trending cleanly and slowly when it is oscillating. The intent is to get responsiveness where it helps and stability where it does not, and the mechanism does behave that way.

The practical benefit shows up in ranging conditions, where a conventional average whipsaws and this one flattens. That reduces the false crossovers that make trend-following systems expensive during sideways stretches, which is where most of their losses accumulate.

The mistake is expecting the adaptation to anticipate. The efficiency ratio is computed from recent price, so the average speeds up after a trend has begun and slows down after a range has established itself. It adapts to conditions rather than to what conditions are becoming.

The adaptive mechanism also introduces its own parameters, including the efficiency lookback and the fast and slow bounds. Those need setting, they interact, and the resulting search space is larger than a simple average's single period, which makes overfitting easier rather than harder.

KAMA FAQs

References

• CMT Association: Technical Analysis Body of Knowledge and Research
• CFA Institute Research and Policy Center: Investment Research
• TA-Lib: Technical Analysis Function Library Documentation
• Perry J. Kaufman, Trading Systems and Methods, the reference source describing the Kaufman Adaptive Moving Average and its Efficiency Ratio.
• SEC Investor.gov: Introduction to Investing