Chande Kroll Stop: Formula, the Two Sets of Settings, and a Daily Bitcoin Example
Tushar Chande and Stanley Kroll’s volatility stop, from the book that introduced it: how its two lines come from the highest high, the lowest low and the average true range, why 10, 1, 9 and 10, 3, 20 both circulate as defaults, and how each behaved on daily Bitcoin.

On this page
The Chande Kroll Stop is a trailing stop built from volatility. Tushar Chande and Stanley Kroll described it in their 1994 book The New Technical Trader. It draws two lines on the price chart: one hangs below the recent highest high by a multiple of the average true range, the other sits above the recent lowest low by the same multiple. A second step keeps each line at its most extreme value of the last few bars, so a stop does not jump back with every new bar.
Chande Kroll Stop (10, 1, 9) on daily Bitcoin

- Price, daily candles (hollow up, filled down)
- Line from the highs: highest of (highest high − 1 × ATR) over 9 bars (solid line)
- Line from the lows: lowest of (lowest low + 1 × ATR) over 9 bars (solid line, second color)
How the Chande Kroll Stop is calculated
LuxAlgo gives the steps, set out in the box below. The first stop from the highs is the highest high of the last p bars minus x average true ranges; the first stop from the lows is the lowest low plus the same amount. Each line then holds its first stop at its extreme over the last q bars: the highest for the line from the highs, the lowest for the line from the lows. TradingView’s help page adds that the indicator is based on the true range, the range of a bar extended to the previous close when there is a gap; see our page on the average true range.
The second step is what sets the Chande Kroll Stop apart from a plain volatility stop. When one bar with a wide range lifts the average true range, the first stop drops at once, but the line keeps the highest first stop of the last q bars, so it stays where it was until that level leaves the window.
How the Chande Kroll line from the highs is built

Which settings: 10, 1, 9 or 10, 3, 20?
Several sets of defaults circulate. LuxAlgo gives 10, 1 and 9 as following the common charting implementation, and notes that the book and the platforms vary in their exact parameters. The open source pandas-ta-classic library uses the same 10, 1 and 9 as its TradingView mode, with Wilder’s smoothing of the true range, and also offers the book’s defaults, 10, 3 and 20 with a simple average of the true range, citing page 95 of The New Technical Trader for them. ProRealTime’s version comes with its own defaults: p = 20, q = 10 and x = 3 (20, 3, 10 in the order used on this page). Check which one your platform uses before comparing levels. TradingView lists the Chande Kroll Stop among its built-in indicators, so you can open a free TradingView account (affiliate link) and compare both settings on your own market.
The multiple x changes the nature of the lines. With x = 3, the line from the highs sits three average ranges below the top, which is where a stop for a long position goes. With x = 1, that line is close to the highs and often above the close. The chart below draws the book’s settings on the same 250 days as the chart under the introduction: the lines sit further from the extremes and cross each other for long stretches.
Chande Kroll Stop (10, 3, 20), the book’s settings, on daily Bitcoin

- Price, daily candles (hollow up, filled down)
- Line from the highs: highest of (highest high − 3 × ATR) over 20 bars (solid line)
- Line from the lows: lowest of (lowest low + 3 × ATR) over 20 bars (solid line, second color)
Which line is the long stop?
Platforms do not label the lines the same way. TradingView’s help page describes the red line as the stop for short positions and the green line as the stop for longs. It reads a close below both lines as a sell and a close above both as a buy, and adds that some traders act when the two lines cross each other. In code, pandas-ta-classic calls the line built from the highs the long stop, while LuxAlgo’s formula names it the short stop. Our chart avoids the problem by naming each line by how it is built: from the highs or from the lows.
The two settings on a real chart
The chart under the introduction shows the 10, 1, 9 lines on daily Bitcoin from February 21 to October 28, 2026. We computed both settings over the same 250 days and counted how often the close crossed each line, without trading anything.
| Settings | Line from the highs against the close, on average | Days the line from the highs was above the line from the lows | Closes crossing below the line from the highs | Closes crossing above the line from the lows |
|---|---|---|---|---|
| 10, 1, 9, Wilder’s average | 3.12% above the close | 250 of 250 | 19 | 18 |
| 10, 3, 20, simple average | 0.81% below the close | 81 of 250 (32.4%) | 7 | 12 |
With 10, 1 and 9 the line from the highs stayed above the line from the lows every day, and the close moved below the line from the highs or above the line from the lows 37 times, about once a week; counting the moves in the other direction too, the close crossed a line 74 times. With 10, 3 and 20 the lines were three ranges away from the extremes and crossed each other: the line from the highs was below the line from the lows on 169 of the 250 days. The close moved below the line from the highs or above the line from the lows 19 times. The wider setting gives fewer, later signals; whether that is better depends on what the stop is for.
The count does not fall evenly as the multiple grows. Keeping p = 10 and q = 9 with Wilder’s average and moving only x, the chart below counts the same two kinds of crossing over the same 250 days.
Closes crossing each line in 250 days, by setting

- Close fell below the line from the highs
- Close rose above the line from the lows
Going from 1 to 1.5 added signals, because the line from the highs dropped from above the close to around it; from 1.5 to 3, the count fell from 40 to 22. The book’s 10, 3, 20 gave 19.
Chande Kroll Stop and Chandelier Exit
The Chandelier Exit, by Charles Le Beau, also places a trailing stop from the period high or low using the average true range and a multiplier, over 22 periods by default according to StockCharts. The Chande Kroll Stop adds the second step: it keeps the most extreme stop of the last q bars, which holds the line in place for a while when volatility jumps. Both work as a trailing stop, not as a forecast.
Does it work?
We have no long-run measurement of the Chande Kroll Stop: our bench tests chart patterns detected by rule against random entries, not indicators (see our methodology). The counts above describe the lines on one market and are not a test. A fair test would compare exits on these lines with a fixed stop on many markets, with costs.
For other volatility and trend stops, see the Parabolic SAR and the Gann HiLo Activator, or the full list of trend indicators.
Frequently asked questions
How do I set up the Chande Kroll Stop indicator?
Choose three numbers: p, the bars for the highest high, lowest low and average true range; x, the multiple of the range; q, the bars over which each stop is held. Charting platforms commonly use 10, 1 and 9; the book’s settings are 10, 3 and 20. A larger x puts the lines further from price.
Who created the Chande Kroll Stop?
Tushar Chande and Stanley Kroll, in their book The New Technical Trader, published in 1994.
What is the difference between the Chande Kroll Stop and the Chandelier Exit?
Both hang a stop from the period high or low by a multiple of the average true range. The Chande Kroll Stop then keeps the most extreme stop over the last q bars, a second smoothing step the Chandelier Exit does not have. See our page on the Chandelier Exit.
How do you read the two lines?
On TradingView’s reading, a close below both lines is bearish and a close above both lines is bullish. Each line can also serve as a trailing stop level. Platforms label the lines differently, so check which line your chart calls the long stop.
