Parabolic SAR

The one that flips

So far every indicator we've covered either signals direction or strength — it tells you what the market is doing. Parabolic SAR does something different: it tells you when to get out. And because it's also stop-and-reverse, getting out on a long signal means entering a short in the same bar.

Murphy's plain definition:

The letters "SAR" stand for "stop and reverse," meaning that the position is reversed when the protective stop is hit. It gets its name from the shape assumed by the trailing stops that tend to curve like a parabola. — John Murphy

Kaufman's sharper philosophical framing, worth reading twice:

The philosophy of the Parabolic System is that time is an enemy. Once a position is entered, it must continue to be profitable or it will be liquidated. — Perry Kaufman, Trading Systems and Methods

That's the whole idea. The stop accelerates toward price over time. If price keeps moving in your favor, the stop chases and locks in profit. If price stalls, the stop catches up and takes you out — even if nothing bad has technically happened. SAR doesn't wait for a reversal signal. Time itself is the signal.

J. Welles Wilder Jr. published SAR in New Concepts in Technical Trading Systems (1978) — same book as RSI, ATR, ADX, and Swing Index. All five indicators share Wilder's α = 1/N smoothing convention and his general "trend-follower's toolkit" philosophy.

The formula

Three pieces:

• SAR: the stop price for the current trade. In uptrends it sits below price (long-trade stop); in downtrends above (short-trade stop).
• EP (Extreme Point): the highest high reached during the current uptrend, or the lowest low during a downtrend. Kaufman: "Using the highest high of the current move when long and lowest low when short keeps the SAR at its highest possible level during an upward move and lowest during a downward move."
• AF (Acceleration Factor): starts at 0.02, increments by 0.02 each time a new EP is made, capped at 0.20.

Kaufman's description of the AF mechanic:

The smoothing constant is called the acceleration factor (AF), and it is initially set to 0.02 at the beginning of each trade. After a day in which a new extreme occurs (a new high when long, or a new low when short) the AF is increased by 0.02… AF cannot be increased above the value 0.20, the equivalent of a 9-day moving average.

The reversal rule

When price touches SAR, you flip. Long becomes short, short becomes long. On the flip:

1. SAR becomes the prior trend's EP (the extreme reached during the just-completed trade).
2. EP resets to the current bar's opposite extreme (current bar's low for a new downtrend, current bar's high for a new uptrend).
3. AF resets to 0.02.

The flip is immediate, unconditional, and gives a natural entry signal for the reversed position. No separate entry rule is needed — SAR is always in the market.

The anti-whipsaw rule

Pure SAR math would let the stop jump into the last 2 bars' range, producing ugly same-day reversals on minor noise. Wilder added a constraint, quoted verbatim from Kaufman:

If long, the SAR may never be greater than the low of today or the prior day. If it is greater than this low, set the SAR to that low value. A reversal will occur on a new intraday low, which penetrates the SAR. For short positions… the SAR may never be below the high of today or the prior day.

This keeps SAR at least a 2-bar-low's distance away until a genuine violation occurs. It doesn't eliminate whipsaws but it filters out the smallest ones.

Play with it

Dots appear below price in uptrends (green, trade-your-long stop) and above price in downtrends (red, trade-your-short stop). Try:

• Up market → green dots track price upward, tightening as new highs accumulate and AF climbs toward 0.20.
• Down market → red dots track price downward, mirror logic.
• Sideways market → reversal count explodes. Every minor swing triggers a flip. Count the reversals; this is SAR's failure mode.
• Crank AF max up to 0.30 → dots tighten faster, more reversals even in trends.
• Drop AF start to 0.01 → dots give more room early in a trade, fewer whipsaws on entry.

What the dots mean

Murphy's practical description:

In an uptrend, the rising dots below the price action (the stop and reverse points) tend to start out slower and then accelerate with the trend. In a downtrend, the same thing happens but in the opposite direction (the dots are above the price action). The SAR numbers are calculated and available to the user for the following day.

That last detail matters: tomorrow's SAR is computed from today's data. The dot you see at bar t is the stop level for bar t+1's trade. Some charting platforms render the current bar's SAR; others render the next bar's. Read your platform's docs.

Strengths — and the one Kaufman really likes

Kaufman on what SAR gets right, verbatim:

A strong point is that the initial stop and reversal is the market extreme high or low rather than a statistically generated point. Also, the SAR does not get closer than the recent 2-day high or low, which prevents a reversal due to noise when the market is moving strongly in a profitable direction.

Translation: SAR doesn't use arbitrary N-day lookbacks or σ-based stops. It anchors to actual price extremes — the highs and lows that real buyers and sellers defended. In a strong trend this is clean and adaptive.

Kaufman's category placement is telling:

Some trailing stops advance, but never retreat, as in Wilder's Parabolic.

Trailing-only-one-way is a useful structural property. Many "trailing stops" actually loosen on pullbacks; SAR doesn't.

Weaknesses — Wilder's own 30% warning

Murphy's bluntest sentence about every SAR-style trend system:

They work well during strong trending periods, which Wilder himself estimates occur only about 30% of the time.

You saw this same statistic in the ADX lesson. It's the pair's shared premise: trend systems are wrong for 70% of market conditions by design. SAR in a trendless period is a money shredder. Our playground above demonstrates this at one click — flip to sideways and count reversals.

Kaufman also flags a subtler weakness with the AF convention:

One weak point may be that AF always begins at 0.02. Although this allows the trade latitude to develop without being stopped out, a market that is moving quickly may be better traded starting with a faster trend.

Meaning: a market that's already strongly trending when you enter gets treated by SAR as if it's just starting — giving back early profit to let the trade "develop" when it's already developed. Some variants (Volker Knapp's 2010 alternative cited by Kaufman) replace the fixed AF schedule with one driven by an efficiency ratio or ATR.

Pair with ADX — the Directional Parabolic System

The single most important SAR upgrade, and the one Wilder himself advocated in the same 1978 book. Murphy:

Directional Movement can be used either as a system on its own or as a filter on the Parabolic or any other trend-following system… by using the Directional Movement as a filter, several of the bad signals in the Parabolic could be avoided.

The combined rule set (from Kaufman's Directional Parabolic System):

• ADX rising AND +DI > −DI → only take SAR LONG flips; ignore SAR shorts.
• ADX rising AND −DI > +DI → only take SAR SHORT flips; ignore SAR longs.
• ADX below 20 → stand aside. Don't trade SAR at all in non-trending regimes.

This is mechanical regime filtering applied to a trailing-stop system. It halves (roughly) the number of trades while dramatically improving the average winner/loser ratio. Murphy's example: "That would have eliminated several whipsaws during the rally phases."

Not a standalone entry signal

SAR generates entries only because it's always in the market. That doesn't make SAR a good entry system — it makes it an exit system that happens to have an entry side effect. Both books treat SAR primarily as:

1. A trailing stop (when you enter via a different system)
2. A regime-dependent always-in system (when gated by ADX / DMI)

Using raw SAR flips as your only signal in unfiltered markets is the beginner trap.

Hidden traps

• Raw SAR in ranges. Universal failure mode. Requires ADX or some regime filter.
• AF over-tuning. Defaults are Wilder's 0.02 / 0.02 / 0.20. Cranking AF max to 0.30+ makes the stop too aggressive; dropping it below 0.10 defeats the "accelerate with the trend" purpose.
• Misreading the dot position. Green dot below price = the stop for the long trade you are currently in. Not an entry signal, not a price target. When price touches it, you flip.
• Expecting SAR to catch reversals. SAR is a following tool — it rides trends and gets flipped out when they fail. It doesn't predict turns; it reacts to them.
• Using SAR without knowing which bar's SAR is shown. Some platforms plot the current bar's SAR; others the next bar's. These lead to different visual interpretations.
• Forgetting the 2-bar constraint. If your implementation lets SAR penetrate the last 2 bars' range, it's not Wilder's SAR — it's a noisier variant.
• Running SAR on low-volatility instruments. Kaufman: "This method requires fairly consistent price swings to produce profits." Flat instruments make SAR whipsaw endlessly.

Quick check

What you now know

• Parabolic SAR (Wilder, 1978) is a trailing-stop indicator that flips position when hit — "Stop And Reverse." Same book as RSI / ATR / ADX / Swing Index.
• Formula: SAR(t+1) = SAR(t) + AF × (EP − SAR(t)). EP = highest high (uptrend) or lowest low (downtrend). AF starts at 0.02, bumps by 0.02 on each new EP, caps at 0.20.
• Reversal logic: when price penetrates SAR, flip. New SAR = prior trend's EP; new EP = current bar's opposite extreme; AF resets to 0.02.
• Anti-whipsaw constraint: SAR cannot jump closer than the past 2 bars' range. Kaufman: "the SAR does not get closer than the recent 2-day high or low."
• Dots below price = long-trade trailing stop; dots above = short-trade trailing stop. When price crosses a dot, flip.
• The "time is an enemy" philosophy (Kaufman): the accelerating stop forces continuous forward progress or takes you out.
• Wilder's ~30% trending estimate applies here too — SAR alone whipsaws in ranges.
• Pair with ADX (Wilder's own Directional Parabolic System) is the canonical upgrade. ADX regime filter + SAR trailing stop = the combo both books recommend.
• AF defaults (0.02 / 0.02 / 0.20) are Wilder's; variants exist (Volker Knapp 2010 replaces fixed AF with an efficiency-ratio-driven version), but deviating from defaults without backtesting is curve-fitting.

Next up in the More Indicators unit: Keltner & Donchian Channels — two non-Bollinger ways to build an envelope around price, one ATR-based (Keltner), one high-low-based (Donchian).