Parabolic SAR: How It Works and When It Fails

J. Welles Wilder introduced the Parabolic SAR in his 1978 book alongside RSI and ATR. It's one of the few indicators designed as a complete "stop and reverse" system — always in the market, always with a defined stop. The dots on the chart make it visually intuitive, and the automatic trailing behavior appeals to traders who struggle with manual stop management. But the indicator has a well-known weakness that backtests quantify clearly.

How Parabolic SAR Works

SAR stands for "Stop and Reverse." The indicator is always either long or short — there's no neutral state. A series of dots appears below price during uptrends and above price during downtrends. When price touches a dot, the system reverses.

The formula uses two variables:

• AF (Acceleration Factor): Starts at 0.02 and increases by 0.02 each time price makes a new extreme (new high in uptrend, new low in downtrend), up to a maximum of 0.20
• EP (Extreme Point): The highest high in an uptrend or lowest low in a downtrend

The SAR value for the next period is calculated as:

SAR(next) = SAR(current) + AF × (EP - SAR(current))

This "parabolic" acceleration is the key feature. As a trend continues, the SAR dots curve closer to price, creating tighter and tighter stops. If the trend is strong, this captures most of the move. If the trend stalls, the tightening stop catches the reversal early.

The Acceleration Factor in Practice

The default AF of 0.02 with a 0.20 maximum means the indicator starts conservatively and accelerates over time. At the start of a new trend, the dots are far from price — giving the move room to develop. As the trend continues and makes new highs (or lows), the dots approach faster.

The maximum AF of 0.20 prevents the dots from getting too close too quickly. At AF = 0.20, the SAR moves toward price at 20% of the distance per bar. This means even the tightest trailing stop still gives some breathing room.

Adjusting these parameters changes the indicator's personality dramatically:

Where SAR Excels: Trending Markets

Parabolic SAR was built for trends. In a strong, sustained move, the indicator performs excellently — capturing the majority of the trend and exiting near the reversal. Our backtest on BTC/USDT daily during the 2020-2021 bull run shows this clearly.

Using default 0.02/0.20 settings, SAR captured 73% of the October 2020 to April 2021 BTC move ($11,000 to $64,000), exiting at approximately $58,500 when the acceleration brought the stop too close during the April top. For a single indicator with no additional filters, this is a strong result.

The same settings on EUR/USD during the sustained downtrend of June-September 2022 captured 82% of the move, demonstrating that SAR's trend-following capability is consistent across asset classes when a clear trend exists.

Where SAR Fails: Ranging Markets

This is SAR's Achilles heel, and it's not subtle. In a sideways market, price bounces between support and resistance, hitting the SAR dots on both sides. Each hit triggers a reversal, and each reversal starts with the AF reset to 0.02, meaning the new stop is far from price. The result is a series of small losses as the indicator whipsaws.

Quantifying this on BTC/USDT daily during the May-September 2023 range ($25,000-$31,000): SAR with default settings produced 14 signals. Win rate: 21%. Profit factor: 0.42. The few small winners were overwhelmed by frequent small losses. This is not a marginal failure — it's a significant capital drain.

This ranging-market weakness is not a flaw that can be fixed with parameter optimization. It's structural. SAR requires price to move directionally to be profitable. When price oscillates, any "stop and reverse" system will whipsaw by definition.

Solutions for the Ranging Problem

The most effective solution is not to use SAR alone. Combine it with a filter that identifies trending vs. ranging conditions:

• ADX filter: Only trade SAR signals when ADX > 25 (trend is present). This eliminates most whipsaw signals. In our BTC daily test, adding an ADX > 25 filter improved SAR's profit factor from 0.98 (barely breakeven over the full period) to 1.41.
• Higher-timeframe confirmation: Only take SAR signals on 4H that agree with the daily trend direction. This reduced trades from 186 to 82 but improved profit factor from 1.12 to 1.51.
• Bollinger Band width: Only trade SAR when Bollinger Bandwidth is above its 20-period average (volatility is expanding, suggesting a trend). Profit factor improved from 0.98 to 1.28.

SAR as a Trailing Stop (Not Entry)

Many experienced traders use SAR not as an entry signal but purely as a trailing stop mechanism. You enter based on another indicator (e.g., RSI, moving average crossover, or breakout) and use SAR to manage the exit. This avoids the whipsaw problem because you're not entering on SAR reversals in ranging markets.

In this configuration, the 0.01/0.10 settings (slow step, low maximum) tend to perform best because they give the trade more room to develop. The standard 0.02/0.20 can tighten too quickly on lower timeframes.

Practical Considerations

SAR cannot handle gaps well. If price gaps past the SAR value, the reversal occurs at the gap's opening price, which may be far from the original stop level. This is relevant for stocks (overnight gaps) and crypto (weekend gaps on some exchanges). On 24/7 crypto markets, this is less of an issue.

SAR also resets the acceleration factor on every reversal, meaning the initial stop after a reversal is always the widest. If you're using SAR as a trailing stop and the trend is young, the first few bars will have a very loose stop. This is by design — it gives the new trend room — but it means your risk per trade varies significantly depending on where you are in the SAR cycle.