Parabolic SAR: The Trailing-Stop Indicator That Whipsaws in Chop
A full walkthrough of the Parabolic SAR indicator: what the dots mean, the formula and the acceleration factor behind them, how to use it as a trailing stop, the settings that actually differ by trading style, the double SAR strategy, and an honest account of why it falls apart in sideways markets.
Parabolic SAR is one of the easiest indicators on any platform to read and one of the easiest to lose money with. A line of dots under price means uptrend, a line of dots over price means downtrend, and the dot itself is a ready-made stop level that tightens as the move runs. That is genuinely useful, and it is also the whole reason the indicator is dangerous: it produces a clean, confident, precise-looking signal in a market that is going absolutely nowhere, and it produces one every two or three bars. StockCharts puts a number on the underlying problem that most guides leave vague, noting that markets trend roughly 30% of the time. The indicator has no way of telling you which 30% you are in. Here is the formula, the settings that actually matter, the pairing that fixes the blind spot, and where the whole thing comes apart.
Quick Answer
Parabolic SAR (Stop And Reverse) is a trend-following indicator by J. Welles Wilder that prints a dot on every bar. Dots below price mean an uptrend, dots above mean a downtrend, and a dot flipping to the other side is the reversal signal. Each dot is calculated as SAR = prior SAR + AF x (EP - prior SAR), where AF starts at 0.02, rises 0.02 per new extreme, and caps at 0.20. Traders mainly use it as a trailing stop. It works in trends and whipsaws badly in sideways markets.
What Is Parabolic SAR?
J. Welles Wilder published it in New Concepts in Technical Trading Systems in 1978, the same book that gave the world RSI, ATR and the Directional Movement system. That authorship matters more than it sounds like it should. Wilder was not designing a standalone indicator, he was designing components of a system, and two of the pieces in that book exist specifically to answer questions the other pieces cannot. Parabolic SAR answers "where does my stop go and when do I turn around." It was never meant to answer "is this a trend." If you have read the guide to trading RSI you have already met the other half of Wilder's toolkit.
The name is two ideas stuck together. SAR is an acronym for Stop And Reverse, which describes the trading system: the dot is your stop, and when price touches it you exit and open a position in the opposite direction on the same bar. Parabolic describes the shape. Because the calculation accelerates each time the trend prints a new extreme, the dots curve in toward price rather than tracking it at a fixed distance.
What you actually see on a chart is three things. Dots below the price bars mean the indicator has the trend as up. Dots above mean down. And when a dot jumps from one side to the other, that is the signal, which in Wilder's original design is simultaneously an exit and an entry. Fidelity's description of the parabolic stop and reverse system is blunt about what that implies: it is an always-in-the-market system, with no flat state. You are long or you are short, and the flip is the only thing that changes which.
Same indicator, same settings, three very different markets
Look at the right third of that diagram and you have the entire argument of this post in one picture. Same indicator, same settings, same instrument. In the trend the dots do a genuinely good job of riding behind price. In the range they flip on almost every bar, and every one of those flips is a stop-out with a spread crossed and a commission paid.
What Is the Parabolic SAR Formula, and How Does It Actually Calculate?
The headline formula fits on one line. For an uptrend:
Uptrend: SAR = Prior SAR + Prior AF x (Prior EP - Prior SAR)
Downtrend: SAR = Prior SAR - Prior AF x (Prior SAR - Prior EP)
EP (Extreme Point) is the highest high reached since the uptrend began, or the lowest low since the downtrend began. AF (Acceleration Factor) starts at 0.02, increases by 0.02 each time a new EP is recorded, and stops at 0.20.
Read what that does rather than what it says. The term in brackets is the distance between the current stop and the best price the trend has managed so far. Each bar, the SAR moves some fraction of that distance closer to the extreme, and the fraction gets bigger every time the trend makes a new high. Early in a move, AF is 0.02, so the stop creeps 2% of the gap per bar and the trade has room. Twenty new highs later, AF has hit its 0.20 ceiling and the stop is eating a fifth of the remaining gap every single bar. The indicator gets impatient as the trend ages. That is a design decision, not a side effect, and it is why Parabolic SAR tends to exit near the end of a big move and to exit constantly when there is no move.
Three pieces of the calculation get left out of most one-line explanations, and they are the ones that explain behaviour people find confusing on a live chart.
Initialization
The formula needs a prior SAR to run, and on the first bar there is not one. Implementations seed it from the prior period's extreme, so the first few dots on a freshly loaded chart depend on where the data window happens to start. Load the same symbol with 200 bars of history and with 2,000, and the early dots can differ. They converge quickly, which is why nobody complains, but it is worth knowing before you go looking for a bug.
Everything resets on a flip
When price crosses the SAR, three things happen at once. The new SAR becomes the EP of the trend that just ended. The AF resets all the way back to 0.02. And a new EP is established from the extreme of the new direction. This is why the first dot after a reversal sits so far from price: you are back at the slowest acceleration with the widest gap. A trader who flips and immediately complains that the stop is miles away has met this reset.
The two-bar constraint
There is a rule that overrides the formula. The SAR is never allowed to sit inside the price range of the current or the previous bar. In an uptrend, if the formula produces a value above the low of either of the last two bars, it gets pulled down to that low instead. In a downtrend it gets pushed up to the higher of the last two highs. Without this clamp the accelerating SAR would regularly place itself in the middle of a candle and trigger an exit on ordinary bar-to-bar noise. It is the one piece of the calculation that exists purely to stop the rest of it misbehaving. StockCharts' ChartSchool entry on Parabolic SAR works through the arithmetic bar by bar if you want to reproduce it in a spreadsheet.
How Do You Use the Parabolic SAR Indicator?
Practically, there are two uses, and only one of them is widely defensible.
The defensible one is the trailing stop. You get into a trade some other way, and once you are in, you move your stop to the current SAR dot and update it every bar. This is the use case almost every source converges on and it plays to the formula's actual strength: an exit level that starts generous and tightens automatically as the move matures, with no discretion required from you and no chance of the stop moving the wrong way. The wider question of when a trailing exit beats a fixed one is covered in the piece on when a trailing stop earns its keep, and the SAR dot is one concrete way of implementing what that post describes.
The less defensible one is taking the flip as an entry signal, which is what Wilder originally intended. In a trending instrument this is fine and occasionally excellent. In everything else it is the source of nearly every complaint anyone has ever had about this indicator, for reasons the next section gets into.
A few mechanical details worth having straight before you put it on a live chart. The dot for a closed bar never moves, but the dot on the bar currently forming does, so a flip you can see at 10:42 may not exist when the bar closes. Waiting for the close is the conservative handling and it costs you part of one bar. The SAR level is a chart level, not a guaranteed fill, which matters on a gapping equity in a way it does not on a liquid future. And the dot only ever moves in the direction of the trade, which is the property that makes it a legitimate stop and separates it from the discretionary stop-widening that kills accounts. If you want the reasoning behind where a stop belongs in the first place rather than how to trail it, that is the subject of the breakdown of structural stop placement.
A trailing stop on a chart with no trend behind it is just a slower way to lose the same money.
Upload the chart screenshot and SnapPChart reads that one image against a fixed rubric, then returns a setup grade, an entry, a stop with the reasoning behind the level, targets, and the reward-to-risk those levels imply. The grade is the part worth arguing with before you commit size.
Grade this chartWhy Does Parabolic SAR Whipsaw in Chop?
Eleven of the twelve reference guides I went through to write this say some version of the same sentence: works in trends, whipsaws in ranges. It is the most universally agreed limitation attached to any common indicator, and it deserves to be stated as a hard fact rather than hedged into a caveat at the bottom of the page.
The mechanism is not mysterious. The indicator is a stop-and-reverse system with no neutral state, so it is structurally required to produce a signal in both directions. Give it a market oscillating in a range and it will dutifully flip long at the top of the range and short at the bottom, which is the exact inverse of what a range trader would do. The acceleration factor makes it worse. Every small push to a new local extreme bumps AF, which pulls the dot in closer, which makes the next ordinary pullback more likely to trigger a flip.
StockCharts is the only source in that set that bothers to quantify the exposure, noting that securities trend roughly 30% of the time. Take that at face value and the arithmetic is uncomfortable: for most of the year, the default behaviour of this indicator on most instruments is to generate signals you should be ignoring. SnapPChart's own ranked list of day trading indicators puts Parabolic SAR in the avoid pile for day traders for exactly this reason, since the opening and closing hours where most intraday volume sits are also where the chop is worst.
The timeframe argument nobody has settled
There is a real disagreement in the published guidance here and flattening it would be dishonest. LiteFinance warns explicitly against M5 and M15 charts, calling them too sensitive to market noise, and recommends H1 and above. CMC Markets, Alchemy Markets, TrendSpider and several forex education sites present scalping and short intraday timeframes as perfectly valid applications, with faster acceleration settings to match. Both camps are describing something real. The lower the timeframe, the higher the ratio of noise to trend in each bar, and the more flips you get per hour. That is arithmetic, not opinion.
What both camps agree on, and what actually decides the question, is the state of the market rather than the number on the timeframe button. A 1-minute chart on a stock that gapped up 18% on earnings and is trending hard on heavy volume is a perfectly good home for a fast SAR. A daily chart on a mega-cap grinding sideways for six weeks is not, at any setting. Timeframe is a proxy for the thing that matters, and it is a noisy proxy. Read the trend directly instead, which is the argument the overview of how indicators and structure fit together makes at greater length.
What Are the Best Parabolic SAR Settings?
The aggregate answer from the top-ranking guides on this question is "it depends on your approach," which is true and useless. Here are concrete numbers, with the tradeoff each one buys, because the settings genuinely do different things and picking one is picking a side.
Two parameters, whatever your platform calls them. The step (MT4 names it Step, TradingView names it Start and Increment) is how much AF rises per new extreme. The maximum is the AF ceiling. Raise either and the dots converge on price faster, meaning tighter stops, earlier exits and more flips. Lower either and you get the opposite.
| Style | Step | Max | What it does | What it costs you |
|---|---|---|---|---|
| Wilder's original | 0.02 | 0.20 | The default on every platform. Tightens at a moderate rate and flips after a normal-sized pullback | Tuned for 1978 daily futures charts, not for a 5-minute chart in the first thirty minutes of the session |
| Position / long-term | 0.01 | 0.10 | Dots converge slowly, so the stop stays far behind price and survives multi-day pullbacks | Gives back a lot at the top. You exit well below the high on every reversal |
| Swing trading | 0.015 to 0.02 | 0.10 to 0.20 | The middle. Survives an ordinary two or three day pullback without flipping | Still flips early on a sharp shakeout that resolves back in your direction the next day |
| Day trading | 0.02 to 0.03 | 0.20 | Reacts faster to intraday reversals, which matters when the trade has to be closed by the bell | Flip count rises sharply. Every extra flip in a range is a real stop-out with real slippage |
| Scalping | 0.03 | 0.20 to 0.30 | Very tight. Dots catch up to price within a handful of bars | Only defensible on an instrument that is genuinely trending right now. Anywhere else it is a shredder |
| Wider max, same step | 0.02 | 0.30 | Same early behaviour, much more aggressive tightening once a trend has run | Cuts extended trends short. The thing you were holding for is the thing this exits you from |
The honest summary of that table: there is no best setting, there is only a choice about which failure you would rather have. Loosen it and you keep more trends but give back more at the top. Tighten it and you lock in more of each move but get shaken out of the ones that were going to work. TradingView's documentation for the Parabolic SAR indicator makes the same point about the absence of a universal answer and recommends testing changes rather than adopting someone else's numbers, which is the right instinct. One caution about that testing: the difference between 0.02 and 0.03 on a handful of charts is mostly luck. If you are going to tune it, tune it over enough trades that the result means something, and tune it on the specific kind of instrument you actually trade.
What Should You Pair Parabolic SAR With, and Why That Specific Pairing?
Every guide says do not trade it alone. Most of them stop there. The reason the advice is right is more useful than the advice, so start with the reason.
Parabolic SAR encodes exactly two things: which direction the calculation currently considers the trend to be running, and how far behind price the exit sits. It has no term for how strong the move is, no term for how long it has lasted, and no volume input at all. A dot below price after a 0.3% drift over two hours and a dot below price in the middle of a 12% squeeze are the same output from the same formula. The indicator cannot distinguish them because nothing in the calculation is capable of distinguishing them.
That is precisely the gap ADX fills, and it is why the ADX pairing shows up in so many guides without the reasoning attached. ADX is Wilder's own trend-strength measure from the same 1978 book, and it is explicitly non-directional: it tells you whether a trend exists and says nothing about which way. Bolt the two together and you have the complete thought. ADX says trend or no trend. SAR says which way and where the exit goes. Each one supplies the variable the other is missing, which is why they were published as parts of one system rather than as two independent indicators.
| Pair with | What it adds | Why SAR needs it | Concrete rule |
|---|---|---|---|
| ADX or Wilder's DMI | Trend strength, the one variable SAR has no opinion about | SAR flips with equal confidence in a strong trend and in a two-point range. ADX rising above roughly 20 to 25 is the filter that separates the two | Take SAR flips only while ADX is rising. Stand down when it is flat and low |
| 200-period or 50-period MA | A single directional bias for the whole session | Half of SAR's flips point against the dominant trend, and those are the ones that reverse back immediately | Take long flips only above the MA, short flips only below it. Ignore the other half |
| MACD | Momentum confirmation with a second, slower clock | SAR reacts to the last bar. MACD reacts to the last dozen. Agreement between two different time constants is a real filter | Enter on a SAR flip only when the MACD line is on the same side of its signal line |
| RSI | Where the flip sits inside the recent range | A bullish flip with RSI already at 78 is a late entry with a converging stop behind it, which is the worst combination this indicator offers | Skip flips that fire into an already-stretched reading. Prefer flips out of the middle of the range |
| Volume | Whether anyone actually participated in the reversal | The formula has no volume term at all. A flip on 40% of average volume is a drift, not a turn | Require above-average volume on the bar that produces the flip |
| Higher-timeframe SAR | A slower version of the indicator as its own filter | The noise that flips a 5-minute SAR usually does not move the hourly one. Disagreement between them is a chop warning | The double SAR strategy below. Trade the fast flips only in the direction of the slow one |
The moving average filter is the cheapest of these to implement and probably the highest value per unit of screen clutter. Plot a 50 or 200 period average, take long flips only above it, short flips only below it, and you have thrown away roughly half the signals without having to think about any of them. The mechanics of choosing the period and reading the slope are in the EMA walkthrough, and the same logic applies whether you use a simple or exponential average. If you want the momentum confirmation instead, the MACD guide covers what a crossover is and is not telling you.
Resist the urge to stack all six. Four indicators that all derive from the same price series mostly agree with each other, which feels like confirmation and is not. Two filters that measure genuinely different things, one for trend existence and one for participation, gets you most of the benefit.
What Is the Double Parabolic SAR Strategy?
Instead of adding a different indicator as the filter, the double SAR approach uses a second, slower copy of the same one. Two variants exist and they solve the same problem from different angles.
Two settings on one chart
Plot a slow SAR (0.01 step, 0.10 max) and a fast one (0.02 or 0.03 step, 0.20 max) on the same timeframe. The slow one defines the trend, since it takes a much larger move to flip it. The fast one times the entries. You take a long only when the slow dots are below price, and you use a fast-SAR flip back to the upside as the trigger after a pullback. The fast SAR also becomes your trailing stop once you are in. When the two disagree, meaning the fast one is flipping repeatedly while the slow one holds, that disagreement is itself the signal: it is what a pullback inside an intact trend looks like, and it is also what chop looks like, which is why this variant still wants a volume or strength check on top.
Two timeframes, same settings
Same idea, expressed through timeframe rather than parameters. Default 0.02 / 0.20 on both, with the higher timeframe (hourly, say) setting direction and the lower one (five-minute) timing the entry. Long only when the hourly dots sit below price, entry on the 5-minute flip in that direction. Some published versions add a MACD crossover on the entry timeframe as a third condition, which is reasonable and also the point at which you should ask whether you are filtering or just collecting agreement.
What this actually buys you is worth being precise about. The double SAR does not fix the indicator's blind spot, because two instruments that cannot measure trend strength still cannot measure trend strength between them. What it does is throw away the counter-trend half of the fast signals, and in a range that means roughly half the whipsaws disappear. Useful. Not a solution. The trades that survive the filter are still taken on an indicator with no view on whether the trend has any force behind it, and the disagreement state between the two SARs is genuinely ambiguous rather than informative. Treat the double SAR as noise reduction and keep a separate read on whether the market is trending at all.
Where the tool I build sits in this is narrow, and worth saying plainly rather than stretching. SnapPChart does not plot Parabolic SAR, does not compute acceleration factors and does not read SAR dots off an uploaded image, and a neutral description of what a single chart read does cover is on the AI chart analysis page. The connection to this post is indirect: understanding why SAR falls apart in chop is really understanding that a precise-looking signal can be produced by a market with no trend in it, and that is exactly the sanity check worth applying to any graded setup, mine included. If a written read says a setup is worth taking and you cannot find the trend context it is describing, the disagreement is information. The same applies when you are reading a chart against the levels you marked yourself.
Use Parabolic SAR as a trailing stop and be very selective about using it as an entry. Default 0.02 / 0.20 is fine, and moving it is a choice about which failure mode you prefer rather than an optimisation. Put a trend-strength read next to it, because that is the one thing the formula structurally cannot give you. And if you cannot say out loud why the instrument in front of you is trending right now, the dots are telling you nothing worth acting on.
Frequently Asked Questions
Does Parabolic SAR repaint, and should you wait for the candle to close?
It does not repaint in the sense people usually mean. Once a bar closes, the dot printed against that bar is fixed and will never move, because the calculation only uses prior-bar inputs. What does move is the dot on the candle that is still forming, and that trips up a lot of people the first time they see it. If price pushes to a new extreme mid-bar the current reading shifts, and a flip that appears at 10:42 can be gone by the time the bar closes at 10:45. So the honest answer is no repainting on history, live movement on the current bar. The practical rule several charting guides give is to treat a flip as real only after the bar closes. That costs you the remainder of one bar on every signal, which on a five-minute chart is a real cost and on a daily chart is basically nothing. Acting intra-bar is a legitimate choice if you accept that some of those flips will un-flip.
Is Parabolic SAR a leading or a lagging indicator?
Lagging. Every input to the current dot comes from bars that have already closed: the prior SAR, the prior acceleration factor, and the extreme point reached so far. Nothing in the formula anticipates anything. Some guides describe the lag as unusually small because the value updates every single bar rather than averaging over a window, and that is fair as far as it goes, but small lag is still lag. The more useful framing is what the indicator is actually blind to. It encodes direction and it encodes a converging exit level, and it encodes nothing at all about conviction. A one percent drift higher over twenty bars and a violent trend move both produce dots below price. The indicator cannot tell you which one you are in, which is the whole reason the pairing advice in this post exists.
What happens if you take every Parabolic SAR signal the way Wilder intended?
You end up permanently in the market, long or short, with no flat periods at all, because a stop-and-reverse system by definition exits one side and enters the other on the same bar. Wilder designed it that way on purpose, and Fidelity still describes it as an always-in-the-market system. In a market that trends for months the always-in behaviour is a feature, since you never miss the start of the next leg. In the ranges that make up most of the trading year it means you are paying spread and commission to flip position every few bars in a market going nowhere. Almost nobody trades it as Wilder specified anymore. The common adaptation is to use the flip as a trailing exit and to require a separate condition before taking the entry on the other side, which turns an always-in system into a sometimes-in one.
Does Parabolic SAR work on stocks, or is it mainly a forex indicator?
It is asset-agnostic. Wilder built it for futures in 1978, most of the tutorials you will find online are written by forex brokers because that is who publishes indicator content, and it calculates identically on equities, futures, indices and crypto. Two things are genuinely different on US equities. First, gaps. The SAR dot assumes you can exit at the level, and a stock that closes at 42.10 with the dot at 41.80 can open at 39.50 on news, which is not a defect in the indicator so much as a reminder that no chart-derived stop is a guaranteed fill. Second, session structure. Equities have an open and a close, and the volatility clustering around both is exactly the environment that produces the fast flip-flop behaviour. Twenty-four-hour markets spread that noise out differently. The indicator does not care which instrument it is plotted on. It cares whether that instrument is trending.
Why is it called Parabolic SAR?
Two separate words doing two separate jobs. SAR stands for Stop And Reverse, which describes the system Wilder wrapped around it: the dot is your stop, and when price hits it you do not just exit, you reverse into the opposite position. Parabolic describes the shape the dots trace on the chart. Because the acceleration factor increases every time the trend prints a new extreme, the gap between price and the dot does not close in a straight line. It closes faster and faster, so the line of dots curves in toward price the way a parabola does. That curve is the indicator's entire personality in one visual. Early in a move it gives the trade room. Late in a move it gives almost none, which is why it tends to take you out near the end of a strong trend and why it takes you out constantly when there is no trend to be near the end of.
This article is for educational and informational purposes only and is not investment, financial or trading advice. The Parabolic SAR formula, acceleration factor defaults, extreme point handling and two-bar constraint described here are the conventional formulations published by Wilder and reproduced by charting platforms and reference sources; the settings quoted by trading style are ranges commonly cited in that literature rather than tested or recommended parameters, and no combination of them is claimed to be profitable. The observation that markets trend roughly 30% of the time is attributed to StockCharts and is a general market characterisation, not a backtested result for any strategy in this post. Nothing here is backtested performance and no results are claimed or implied. Indicator-based stops are chart levels, not guaranteed fills, and gaps, halts and thin liquidity can execute well beyond them. Day trading and active trading carry a substantial risk of loss and are not suitable for every investor. SnapPChart grades a static chart screenshot you upload and returns a setup grade, entry, stop, targets and reasoning for that single image; it does not plot or calculate Parabolic SAR, does not track indicator values, does not scan the market, and does not track your account, positions or P&L. Do your own analysis, size positions so that being wrong is survivable, and consider speaking to a licensed financial professional about your own circumstances before trading.
Writes about AI-assisted day trading, technical analysis, and the systems traders actually use to stay disciplined.
The dots tell you where. They never tell you whether.
Parabolic SAR will hand you a stop level in any market, including the ones you should not be trading. Upload a chart screenshot and SnapPChart reads that single image against a fixed rubric: a setup grade, an entry, a stop with the reasoning behind the level, targets, and the reward-to-risk those levels imply. It does not plot SAR. It does give you a written second opinion on whether the trend context is there before you commit.