How does Parabolic SAR work in forex?

Direct answer

Parabolic SAR (Stop and Reverse) is a technical indicator that produces a sequence of calculated levels—often shown as dots—that tend to “hug” the recent price trend. It is designed to track whether price is, at a given moment, moving more strongly upward or downward relative to the indicator’s current dots. When price crosses the dots, the indicator flips its assumed direction and starts building a new sequence from that point.

In forex terms, Parabolic SAR does not require anything special about currency markets. It uses price highs and lows (the “extreme” points of a bar) and a small set of parameters to generate a step-by-step output. The key point is that the indicator’s values are entirely determined by the chosen settings and the historical high/low data used in the calculation—there is no built-in prediction.

A simple model of what it outputs

Parabolic SAR outputs one value per bar (or per time step) that is drawn as dots on the chart. Conceptually, that dot value acts like a trailing reference level:

• In an upward-trend phase, SAR is typically placed below price and moves upward as new highs appear.
• In a downward-trend phase, SAR is typically placed above price and moves downward as new lows appear.

A flip in the indicator happens when the current SAR level is breached by price in the opposite direction. After a flip, the SAR sequence is re-initialized using new starting conditions tied to the reversal bar.

Because SAR uses an iterative update, the dots are not independent from one another. Each new SAR point depends on the previous SAR point, the most recent extreme price in the current assumed trend, and the current acceleration settings.

Mechanics: inputs, calculations, and sequence

To understand how Parabolic SAR works, it helps to separate the stable mechanics from the variable choices.

Inputs (what you need)

1. High and low prices by bar: For each time step, the indicator needs the high and low used to identify extremes.
2. Starting acceleration factor (AF): This controls how quickly SAR can move toward price.
3. Acceleration step: This determines how AF increases as the trend continues.
4. Maximum acceleration factor: AF is capped so the indicator does not accelerate indefinitely.

Different charting platforms may use different default values for AF, the step, and the cap. The mechanics remain the same: AF is updated and applied in an iterative SAR formula.

The sequence (how one SAR value is computed)

A high-level description of the iteration is:

1. Maintain a trend direction state (uptrend-phase or downtrend-phase).
2. Track the extreme price in the current state: For an uptrend-phase, this is generally the highest high since the last flip; for a downtrend-phase, it is generally the lowest low since the last flip.
3. Update the acceleration factor:
   • AF starts at the chosen initial value.
   • While new extremes are made in the same direction, AF increases by the chosen acceleration step.
   • AF increases until it reaches the maximum cap.
4. Compute the new SAR level from:
   • the previous SAR level,
   • the current extreme price,
   • and the current AF.
5. Apply placement rules to prevent SAR from being “inside” recent bars in an unrealistic way. In practice, many implementations use constraints based on recent highs/lows (for example, ensuring SAR does not cross beyond relevant prior extremes during the current phase).
6. Check for a flip condition:
   • In an uptrend-phase, if price moves below the SAR level, the indicator flips to a downtrend-phase.
   • In a downtrend-phase, if price moves above the SAR level, the indicator flips to an uptrend-phase.
7. Re-initialize after a flip so the next SAR values follow the new assumed direction.

A concrete worked-style example (with assumptions)

No live forex data is required to understand the flow. Suppose you are at a time step where the indicator is in an uptrend-phase.

• You have a previous SAR value from the prior bar: SAR_prev.
• You have an extreme price EP that represents the highest high seen since the last flip.
• You have a current acceleration factor AF.

Under a common formulation, the next SAR would be computed using an update like:

• SAR_next = SAR_prev + AF × (EP − SAR_prev)

Then, placement constraints are applied using recent highs/lows so the SAR is positioned consistently relative to price during the uptrend-phase.

Finally, you check the flip condition:

• If the current bar’s low is below SAR_next (i.e., price crosses under SAR), the indicator flips to a downtrend-phase.

This example shows the iterative logic without committing to any specific platform’s exact placement rule details, because implementations can differ in those constraints.

Material limitations and failure modes

Parabolic SAR can be informative, but it is not reliable as a standalone, predictive signal. Several limitations are material—meaning they can materially change how it behaves.

1) Whipsaws in sideways or choppy conditions

If price repeatedly crosses the SAR dots due to small oscillations, SAR will flip direction often. This produces a “whipsaw” effect: the indicator changes its phase frequently, even if there is no sustained directional movement.

2) Parameter sensitivity (AF and caps)

The acceleration factor settings directly affect how quickly SAR moves. If AF increases too aggressively (or the max cap is high), SAR can react faster to price swings and flip more often. If AF increases too slowly (or the step is small), SAR can lag behind turning points.

Because of this, two users using different AF settings on the same chart can get different flip points.

3) Discrepancies across implementations

Different charting systems can handle the “placement constraints” and the exact flip initialization differently, especially around the bar where the reversal is detected. That means you can see slightly different SAR dot locations even when using the same AF/step/cap values.

4) Costs and execution do not map from the indicator

Even though SAR flips when price crosses it, the real-world outcome for any trading activity depends on factors that the indicator does not include, such as transaction costs, bid/ask effects, and execution timing. Historical indicator behavior also does not establish future relationships.

5) Data assumptions: bar definition matters

SAR is calculated from high/low per bar. If you change the timeframe (for example, from 1-hour bars to 15-minute bars), the high/low sequence changes and the SAR output changes accordingly.