In one sentence: the efficiency ratio measures how much of a market's movement actually went somewhere. It divides the net price change over a set number of bars by the total distance price travelled bar by bar to get there, which gives a number between 0 and 1. A reading near 1 means price moved in something close to a straight line; a reading near 0 means price covered a lot of ground and finished close to where it started.
The measure was introduced by Perry Kaufman in his 1995 book Smarter Trading, where it drives his adaptive moving average (KAMA). The formula is easy to find. What is hard to find is what real readings look like: how low they go, how fast they change, and how often two timeframes on the same asset disagree. This page gives the exact calculation and a worked example, then follows dated readings from MoonWire's weekly Bitcoin, Ethereum and Solana chart reads.
The formula
Over a lookback of N bars, using closing prices:
- Direction (net change) = the absolute difference between the latest close and the close N bars earlier
- Volatility (path) = the sum of the absolute bar-to-bar changes across those N bars
- Efficiency ratio = Direction / Volatility
Because the net change can never be larger than the path taken to produce it, the ratio is bounded between 0 and 1. Our weekly reads compute it over a 10-bar lookback on each timeframe (ten bar-to-bar changes, spanning the last 11 closes), the same 10-period lookback commonly used as KAMA's default.
A worked example. Take six closes: 100, 102, 101, 103, 104, 106. The net change is 6. The bar-to-bar moves are 2, 1, 2, 1 and 2, a path of 8. The efficiency ratio is 6 / 8 = 0.75: a trend with some retracement along the way.
Now take 100, 105, 100, 105, 100, 101. The net change is 1, but the path is 5 + 5 + 5 + 5 + 1 = 21. The ratio is 1 / 21 = 0.05.
The second series moved much further than the first and got almost nowhere. That is the whole idea: the efficiency ratio does not measure how much a market moves, it measures how much of that movement survives.
Three things the number does not tell you
It has no direction. The net change is taken as an absolute value, so a clean rally and a clean sell-off both score near 1. Direction has to come from somewhere else. In our reads it comes from where price closes relative to its 20-period EMA (a close within 0.3% of the average reads sideways); the efficiency ratio supplies only the strength label - strong at 0.50 and above, moderate from 0.30 to 0.50, weak below 0.30.
It is not volatility. A low reading does not mean a quiet market. In the 2026-07-28 Solana read, the 3-day efficiency ratio was 0.08, the lowest in our three-asset coverage that week, while the 3D average true range ran about 7.7% of price - the widest ATR-to-price ratio of the three. A great deal of movement, very little of it net.
It is not a forecast. A reading describes the last N bars. It says nothing about whether the next bar extends that path.
Reading 1: an "uptrend" with almost no travel
The clearest teaching case in our corpus is the 2026-08-04 Bitcoin read. That week the 3D trend label flipped from down to up - and the 3D efficiency ratio fell to 0.029, the lowest of all nine timeframe readings we computed across the three assets. The next lowest was 0.12.
How can a trend read "up" on 0.029? Because the two labels answer different questions. Price closed 0.46% above its 3D 20-period EMA, which is enough to set the direction. The efficiency ratio said the ten-bar path behind that close had almost entirely cancelled itself out, so strength read weak. The same read found the entire 3D band between support and resistance, 2,189.72 points, was narrower than a single average 3-day bar (ATR 3,008.22).
Bitcoin's 3D reading had also travelled a long way in three weeks: 0.15 on 2026-07-21, 0.35 on 2026-07-28 (when the strength label moved from weak to moderate), then 0.029.
Ethereum printed the same shape earlier. The 2026-07-14 Ethereum read described its 3D chart as "almost perfectly directionless" at an efficiency ratio of 0.029, with the recent range narrower than one average bar - while its monthly reading sat at 0.70.
The practical lesson: when a trend label changes, look at the efficiency ratio before reading the change as momentum. A flip on a near-zero reading is a close crossing an average, not a move.
Reading 2: one asset, both extremes, same week
On 2026-07-28, Solana registered both ends of the scale at once. Its monthly efficiency ratio reached 1.00, the metric's ceiling, while its 3D reading dropped to 0.08.
On a 10-bar lookback, a monthly reading that rounds to 1.00 means almost every one of the last ten monthly moves stepped the same way - down, in this case, with price about 42% below the monthly 20-EMA. The 0.08 on the 3D chart means the near-term tape was a grind rather than a slide. Neither number is wrong. They measure different stretches of time: ten months against roughly a month.
The monthly reading was also persistent. It printed 0.98 on 2026-07-14, 0.94 the following week, 1.00 on 07-28, 0.985 on 2026-08-04, 0.9437 on 2026-08-11 and 0.9575 on 2026-08-18. Over the same stretch the 3D reading moved between 0.05 and 0.47. A single "efficiency ratio" for an asset is incomplete; the answer depends on the timeframe you ask it on.
Reading 3: a full board on one date
The 2026-08-04 reads give a complete nine-cell snapshot - three assets, three timeframes, the same day:
| Asset | 3D | Weekly | Monthly |
|---|---|---|---|
| Bitcoin | 0.03 | 0.37 | 0.61 |
| Ethereum | 0.12 | 0.15 | 0.63 |
| Solana | 0.47 | 0.21 | 0.985 |
Two patterns show up that a single reading hides. The monthly column was the most directional for all three assets. And the 3D column ran from 0.03 to 0.47 on the same day, even though all three assets carried an identical direction map that week - 3D up, weekly down, monthly down. Same labels, very different amounts of travel behind the near-term turn.
How fast it moves, and why
The ratio can swing hard in a single week, and it can reverse the next:
| Read | Timeframe | Change on the week | Strength label |
|---|---|---|---|
| Bitcoin, 2026-08-04 | 3D | 0.35 to 0.029 | moderate to weak |
| Ethereum, 2026-08-04 | 3D | 0.44 to 0.12 | moderate to weak |
| Solana, 2026-08-04 | 3D | 0.08 to 0.47 | weak to moderate |
| Solana, 2026-08-11 | 3D | 0.4697 to 0.0882 | moderate to weak |
| Solana, 2026-08-11 | Weekly | 0.2065 to 0.3373 | weak to moderate |
| Solana, 2026-08-18 | Weekly | 0.3373 to 0.1552 | moderate to weak |
The mechanism is in the formula. The numerator depends on only two closes, the first and last in the window, while the denominator adds up every step between them. When the window rolls forward, a large bar can drop out of the start, or price can come back toward where the window now begins, and the net change collapses while the path barely shrinks. So a small change in price can arrive with a large change in the ratio. Solana's 2026-08-04 read is an example: price rose just 0.52% on the week while the 3D efficiency ratio went from 0.08 to 0.47 - and one week later it was back at 0.0882. The 2026-08-18 drop on Solana's weekly chart, a 54.0% decline, was the largest efficiency-ratio decline of any cell on our board that week and the only one that week that changed a trend-strength label.
Using it when you read a chart
- Pair it with a direction measure. It tells you how clean a move was, never which way it went.
- Read it per timeframe. A 0.03 on the 3D chart and a 0.70 on the monthly, as Ethereum showed on 2026-07-14, describe a pause inside a larger one-way move, not a contradiction.
- Watch the reading under an unchanged label. Directionality can drain out of a trend while its direction label holds, as it did on Solana's weekly chart on 2026-08-18.
- Do not over-read one week. A reading that jumps can fall straight back, as Solana's 3D did between 2026-08-04 and 2026-08-11.
- Know the lookback. Ten bars on a monthly chart is ten months; on a 3-day chart it is about a month.
- Know where it came from. In Kaufman's adaptive moving average, the ratio sets how quickly the average follows price: high readings let it track closely, low readings slow it down so it ignores chop.
Glossary
- Efficiency ratio (ER) - net price change over N bars divided by the sum of absolute bar-to-bar changes over the same bars; ranges from 0 to 1.
- Direction / net change - the absolute distance between the first and last close in the window.
- Volatility / path - the total distance travelled close to close inside the window.
- Lookback (N) - the number of bars in the window; 10 in our reads.
- KAMA - Kaufman's Adaptive Moving Average, which uses the efficiency ratio to speed up or slow down its smoothing.
- Trend strength - in our reads, the label derived from the efficiency ratio: strong (0.50 and above), moderate (0.30 to 0.50), weak (below 0.30).
Every reading on this page is quoted from a dated MoonWire chart read linked inline. These are descriptive records of what our scans reported on the date shown. Nothing here is investment advice, and none of it is a recommendation to transact in any asset.
