Parabolic SAR: Wilder's Formula, Settings, and What the Stop and Reverse Rules Did on Real Prices
J. Welles Wilder's stop and reverse system from its documented formula: how the acceleration factor tightens the stop, what 0.02 and 0.20 change, and what the rules did on daily Bitcoin and the S&P 500 ETF, with every trade of the last four months of Bitcoin.

On this page
The Parabolic SAR is a trailing stop that also tells you which side to be on. It places one dot per bar below price in an uptrend and above price in a downtrend, and the dot moves a little closer to price every bar, faster as the trend makes new extremes. When price touches the dot, the position is stopped and reversed: SAR stands for stop and reverse. J. Welles Wilder introduced it in his 1978 book New Concepts in Technical Trading Systems, with an acceleration factor that starts at 0.02 and rises by 0.02 up to 0.20. Below we compute it on real daily prices and run it as Wilder designed it, always in the market, on about two years of daily Bitcoin and S&P 500 ETF prices.
Parabolic SAR on daily Bitcoin, drawn as a stepped line (dots on most platforms)

- Price, daily candles (hollow up, filled down)
- SAR below price: long, stop under the bar (solid line)
- SAR above price: short, stop over the bar (solid line, second color)
How the Parabolic SAR is calculated
Each bar's SAR starts from the previous one and moves toward the extreme point (EP), the highest high of the current uptrend or the lowest low of the current downtrend. The acceleration factor (AF) sets how much of the gap is closed each bar. It starts at the step, 0.02, and adds one step every time the trend makes a new extreme, until it reaches the maximum, 0.20. So a young trend gets a loose stop and a mature, accelerating trend gets a tight one: that is the parabola in the name.
Two rules keep the dot honest. In an uptrend the SAR is never placed above the lows of the two previous bars, and in a downtrend never below their highs, as StockCharts documents Wilder's method. And on a reversal the new SAR starts at the old extreme point, which is why the first dot of a new trend is often far from price.
How the Parabolic SAR trails, then stops and reverses

- EP
- extreme point, the highest high of the uptrend or the lowest low of the downtrend
Read the drawing from left to right: the gap between price and dot shrinks faster as the trend ages, the dot is touched on the first real pullback, and the new short starts with a wide stop at the old high. That wide first stop is what makes the next reversal costly when the market only pauses.
How to read the dots
Most platforms draw the SAR as dots; our chart draws the same values as a stepped line, one step per bar, below price when long and above it when short.
| What you see | What it means |
|---|---|
| Dots below the bars | The system is long; the dot is the stop for the next bar |
| Dots above the bars | The system is short; the dot is the stop |
| Dots switch sides | Price touched the stop: the position was reversed during that bar |
| Dots closing in fast | The trend made new extremes and the AF is near its 0.20 cap |
| Dots flipping every few bars | No trend: the system is being whipsawed |
Settings: 0.02 and 0.20, or slower
The step and the maximum are the only inputs. A smaller step and a lower maximum let the stop trail further behind, so it flips less often; a larger step tightens it. Wilder's defaults are a step of 0.02 and a maximum of 0.20, which StockCharts also uses. We ran both the default and a slower 0.01 and 0.10 on our two daily series as a stop and reverse system, filled at the stop level (or at the open when a bar gapped through it), starting flat after 50 bars, before costs.
| Market and setting | From | Trades | Winning closed trades | Result | Buy and hold over the same period |
|---|---|---|---|---|---|
| Bitcoin, 0.02 and 0.20 | May 5, 2025 | 41 | 22 of 40 | +57.24% | −13.60% |
| Bitcoin, 0.01 and 0.10 | May 29, 2025 | 27 | 7 of 26 | −12.70% | −23.81% |
| S&P 500 ETF, 0.02 and 0.20 | September 3, 2024 | 54 | 18 of 53 | −9.48% | +43.34% |
| S&P 500 ETF, 0.01 and 0.10 | September 6, 2024 | 33 | 13 of 32 | +7.05% | +44.87% |
Stop and reverse against buy and hold, by setting

- Stop and reverse
- Buy and hold
Results include the last position, still open on October 28, 2026 (a short on Bitcoin with both settings, a long entered on the last bar with the default on the ETF, and a short with the slower setting). The ETF prices are dividend-adjusted. On Bitcoin, which fell over the period, a system that could go short beat holding it with the default setting. On the ETF, which rose steadily, the same system lost money while holding gained more than 40%: pullbacks kept reversing it into a short. The quick trades are where the losses sit: with the default setting, all 7 Bitcoin trades and all 15 ETF trades that lasted five bars or less lost money. The slower setting traded less; it lost 12.70% on Bitcoin but turned the ETF loss into a 7.05% gain, still far below holding. Two markets over about two years is an illustration of the trade-off, not a measure of which setting works.
The rules on the chart above
The chart shows daily Bitcoin from July 1 to October 28, 2026, with the default setting. The run starts long before the window, so the position on July 1 is a short opened on June 23 at 63,261.82; it was stopped and reversed on July 4 at 62,979.86, a gain of 0.45%. Inside the window the system then reversed 11 more times, for 12 positions from July 4. The bars show the result of each of the 13 positions, in order; the list gives the dates.
Each Parabolic SAR position on Bitcoin, June to October 2026

- long from July 4, closed July 24 at 64,104.14: +1.79%;
- short July 24 to August 5: −0.71%; long August 5 to 13: −2.05%; short August 13 to 18: −2.76%;
- long from August 18 at 64,986.26 to September 1 at 76,547.48: +17.79%, the August rally;
- short September 1 to 3: −6.04%; long September 3 to 10: −5.20%; short September 10 to 18: −4.21%;
- long September 18 to October 7: +3.53%; short October 7 to 21: −0.92%; long October 21 to 24: −4.41%;
- short from October 24 at 80,099.80, still open on October 28 at a close of 81,343.50: −1.55%.
Three of the eleven trades opened and closed inside the window won, and one of them, the August rally, paid for most of the losses: that is how a stop and reverse system behaves. In September, with Bitcoin closing between about 75,600 and 86,600, the dots flipped on September 1, 3 and 10, and each of those three trades lost between 4.2% and 6.1%.
Used that way, the SAR is one kind of trailing stop: a fixed stop stays where it was set, a trailing stop follows the price up and only ever tightens. The SAR adds the acceleration, so it tightens faster the longer the trend runs.
Fixed stop versus trailing stop

Does it work?
We have no long-run measurement of the Parabolic SAR: our bench tests chart patterns detected by rule against random entries, not indicators (see our methodology). The trades above cover two markets over about two years, worked through to show how the rules behave, before costs; they are not a test. Wilder, as StockCharts reports him, expected the system to work best with trending securities, which he estimated occur roughly 30% of the time. The results above fit that: the system made its money in one or two strong moves and gave back in the ranges.
Wilder described his trend strength measure in the same book, the Directional Movement Index and ADX. For other trailing stops and trend filters, see the Chande Kroll stop, the moving average, or the full list of trend indicators. Not to be confused with the parabolic curve pattern, a chart shape of accelerating prices.
Frequently asked questions
What is the Parabolic SAR used for?
As a trailing stop that tightens as a trend ages, and in Wilder's original form as a stop and reverse system that is always long or short. The dots are often used to place or trail a stop rather than to reverse.
What are the best Parabolic SAR settings?
Wilder's defaults are a step of 0.02 and a maximum of 0.20. A lower step and maximum (for example 0.01 and 0.10) trail further away and flip less often. On our two daily series the slower setting traded less, did worse than the default on Bitcoin and better on the S&P 500 ETF; no setting fits every market, and only a test on your market can tell.
What is the Parabolic SAR success rate?
There is no single number. With the default setting, run as a stop and reverse system since May 2025 on Bitcoin and September 2024 on the ETF, 22 of 40 closed trades won on daily Bitcoin and 18 of 53 on the S&P 500 ETF. The result depends far more on whether the market trends than on the share of winners: a few long trends pay for many small losses.
Is the Parabolic SAR a good stop loss?
It is a reasonable trailing stop in a trend because it moves only in the direction of the trade and speeds up as the trend extends. In a range it sits close to price and gets hit often. Some traders place the stop at the SAR value only after the trade has moved in their favor; we have not tested that.
Should the Parabolic SAR be used with ADX?
Wilder described both in the same book, and a trend strength filter such as ADX is the usual way to keep the SAR out of ranges. We have not measured the combination.
