What is Kaufman Efficiency Ratio, and what does “settings” mean?
Kaufman Efficiency Ratio (often shortened to KER) is an indicator that describes how “efficient” price movement is in terms of direction. In simple terms, it compares the absolute distance between the start and end of a chosen price window with the total absolute movement inside that same window.
When people ask “How do settings change KER?”, they usually mean the inputs used to calculate it. The most common setting is the lookback length (how many bars are included). Some platforms also let you choose which price series to use (for example, close versus another source) or how the indicator handles missing data. Those choices change the underlying calculation, so the resulting KER values can change even if the formula name is the same.
Mechanism: how the calculation responds to lookback and price inputs
KER is driven by two quantities over the same window:
- Net movement: the absolute change from the first price in the window to the last price.
- Total movement: the sum of absolute changes between consecutive prices within the window.
The ratio becomes larger when price travels farther in one net direction relative to the amount of back-and-forth movement. It becomes smaller when the market is choppy, because total movement rises even if net movement is limited.
Lookback length (window size)
- Shorter lookback: the window covers fewer bars, so net movement and total movement react more quickly to recent changes. This typically increases responsiveness but can make KER more sensitive to transient noise.
- Longer lookback: the window smooths over more history. KER can become less reactive to short-lived swings, which may reduce noise sensitivity but can increase lag when conditions change.
Choice of price series
If a platform lets you base KER on different inputs (for example, using close instead of another price), the indicator’s net and total movements change. That can alter whether price appears to be “more directional” over the window. This is not a mystery: the indicator is only as good as the series you feed into it.
Evidence through reasoning: expected behavior in different market conditions
Because KER is a ratio of net movement to total movement, you can anticipate how settings change its character without needing live data.
Trending conditions
In a sustained move, consecutive price changes tend to align directionally. Net movement grows, and total movement grows in a way that is more consistent with net movement. With a shorter window, KER may rise faster as the most recent candles fit the directional pattern. With a longer window, that rise may be slower but more stable.
Range or choppy conditions
In a sideways or oscillating environment, price often reverses. Net movement stays limited, while total movement increases due to back-and-forth. With a shorter window, those reversals affect the calculation quickly, so KER may fall more rapidly. With a longer window, reversals are averaged across more history, so changes can appear slower.
Regime shifts and lag
If the market changes from choppy to trending (or vice versa), any lookback window is a compromise. A longer window averages across old behavior, so KER can lag the shift. A shorter window can react quickly, but it may also swing during brief, non-persistent moves.
Limitations and failure modes to verify
- Implementation differences: two charts labeled “KER” may not match if they use different price inputs, different window handling, or different smoothing/normalization details. Always confirm the exact formula and input assumptions on your platform.
- Sensitivity trade-off: tuning toward faster reaction (shorter lookback) often increases exposure to noise; tuning toward stability (longer lookback) often increases lag after regime changes. There is no universal “best” setting.
- Non-predictive use: KER describes movement efficiency over a window. It does not inherently guarantee future direction, and historical patterns do not establish future results.
- Calculation edge cases: during low-volatility periods, total movement can be small relative to net movement, which can make the indicator behave differently across implementations. Small numerical or data-handling differences can matter.