Parabolic SAR — the trailing-stop indicator
The answer
Parabolic SAR is a trailing stop drawn on the chart as a line of dots. It does not predict anything. It moves a stop closer to price every day a trend continues, and it flips to the other side of price when the trend fails.
Why this costs you money
The dots look like signals. They are not signals. They are the output of a rule you could apply with a ruler, and the rule has one input: how long the current move has lasted.
Here is how the loss happens. A stock has been drifting sideways for 3 weeks. The dots flip from above price to below. You buy, because the tool that told you to buy is on the screen and it looks confident. Four days later the stock comes back to the middle of the same range. The dots flip again. You sell at a small loss. Then they flip once more. In a 3 week range you can take 4 or 5 losing trades, none of them large, all of them real.
Small losses do not feel like a problem. Count them. Five trades at 1.2% each, with brokerage and taxes on every entry and exit, is a 7% hole in an account where the stock finished exactly where it started. The stock did nothing. You lost 7%.
This is the standard failure of every trailing indicator, and Parabolic SAR has the sharpest version of it, because the rule forces the stop to keep tightening whether or not the trend is real.
How it works
SAR stands for "stop and reverse". J. Welles Wilder Jr. introduced it in New Concepts in Technical Trading Systems in 1978, in the same book that introduced ATR and ADX.
The idea is simple. When you are long, you need a stop. As the trade goes your way, you would like the stop to follow. Wilder wrote a rule that moves the stop automatically, and moves it faster the longer the trend runs.
Three things go into each new dot:
- The extreme point. The highest high reached so far in this up move (or the lowest low in a down move).
- The current stop level, which is yesterday's dot.
- The acceleration factor, a small number that starts at 0.02, increases by 0.02 every time the trend makes a new extreme, and stops at 0.20.
Each day the dot moves part of the way from where it was towards the extreme point. The size of that step is set by the acceleration factor. Early in a trend the dot creeps. Later in a trend, after many new highs, the dot closes in quickly. That is where the word "parabolic" comes from. The path of the dots curves.
When price touches the dot, 2 things happen at once. The position is closed, and the indicator flips to the other side of price and starts a new trail in the opposite direction. That is the "reverse" half of the name.
The important sentence is this one: the only information in a Parabolic SAR dot is how far price has travelled and how long it has been travelling. There is no volume in it. There is no comparison to any other stock. There is no measure of whether the trend is strong or weak. It is a distance rule.
What it tells you, and what it does not
It tells you where a mechanical trailing stop would sit today, given the move so far. That is genuinely useful. Most people move their stop by feeling, which means they move it when they are frightened and leave it when they are comfortable. A rule removes that.
It does not tell you whether a trend exists. This is the whole problem. The indicator has no idea whether it is in a trend or a range, and it produces exactly the same confident dots in both.
It does not tell you the trend has ended. A flip means price touched a level that a distance rule placed there. In a strong trend, a normal 3 day pause is enough to touch a tightened dot. You are removed, the trend resumes without you, and the indicator flips back 2 days later — higher than where it took you out.
It does not tell you anything about size. A dot 2% below price on a quiet stock and a dot 2% below price on a volatile one are not the same risk. Parabolic SAR does not scale to the instrument. ATR does, and that is the reason Supertrend was built later.
The decision rule
Use Parabolic SAR only as an exit rule, and only when a separate measure says a trend exists.
If ADX is above 25 and rising, a SAR flip against your position is worth acting on. If ADX is below 20, ignore every flip, because the indicator is measuring noise and reporting it as direction.
Never use a SAR flip as an entry on its own. An exit rule and an entry rule are different problems. Being wrong about an exit costs you some of a profit. Being wrong about an entry costs you capital.
ADX is a trend-strength gauge that runs from 0 to 100 and says nothing about direction. It is covered in article 10 of this cluster. The pairing matters because ADX answers the one question SAR cannot ask.
Try this now
This takes about 5 minutes and it will settle the question for you better than any argument.
- Open a daily chart of 1 stock you actually hold. Set the range to 1 year.
- Add Parabolic SAR with the default settings (step 0.02, maximum 0.20). Every charting app has it under indicators.
- Count every flip in that year — every point where the dots crossed from one side of price to the other. Write the number down.
- Now go back through them and mark each flip as one of 2 kinds. Followed through means price moved at least 3% in the flip's direction before the next flip. Whipsaw means the dots flipped back within about 10 trading days without that move.
- Divide. Write down the fraction that followed through.
What you should see. On a typical liquid stock over a year you will find somewhere between 15 and 30 flips, and on most charts fewer than half will have followed through. The whipsaws will cluster. You will be able to see, with your eye, the sideways stretches where 6 flips happened inside 2 months and none of them meant anything.
Now do 1 more thing. Look at the 2 or 3 largest sustained moves on the chart. Check whether SAR kept you in for the whole move or removed you halfway. This is the honest trade-off in one picture: the indicator that captures a big trend is the same indicator that shreds you in the range before it.
Three real cases
1. NIFTY 50, January to June 2022 — the range that eats trailing stops The Indian index spent the first half of 2022 moving between roughly 15,200 and 18,100 without establishing a lasting direction. It fell, recovered, fell again and recovered again. On a daily chart, a default-setting Parabolic SAR flipped repeatedly through this period. The mechanism is not mysterious: each recovery made a new local extreme, the acceleration factor pulled the dots in, and the next pullback touched them. This is the environment the indicator is worst in, and it lasted 6 months.
2. S&P 500, January to October 2022 — the trend it was built for The US index fell through most of 2022 in a series of declines and sharp counter-rallies. A trailing stop that flipped short early in the year and stayed short through the summer would have captured a large part of the fall. But the same chart contains the counter-rally from mid-June to mid-August 2022, when the index rose roughly 17%, which was more than enough to flip a tightened SAR back to long near the top of that rally. The indicator caught the trend and also handed back a meaningful part of it. Both statements are true of the same year.
3. J. Welles Wilder Jr., 1978 — the source, and what he actually said Wilder published Parabolic SAR alongside ADX in the same book. That ordering is the case study. He did not present SAR as a standalone system. He presented a directional-movement framework for deciding whether a trend existed, and a parabolic stop for managing a position once you believed it did. Almost every modern chart shows the second tool without the first. The failure most traders blame on the indicator is a failure to use the book.
The question that resolves it
A novice looks at a SAR flip and asks: is this a buy or a sell?
An expert looks at the same flip and asks: is price currently in a range or a trend?
The flip carries no information about that. It looks identical either way. Every argument about whether Parabolic SAR "works" is really an argument about a question the indicator never asks, and the answer to that question is available from the chart itself in 5 seconds: are the highs and lows of the last 2 months rising, falling, or overlapping?
What would make this wrong
If Parabolic SAR contained real predictive information, then flips would be followed by moves in their direction more often than a coin flip, across many stocks and many years. Run the count in the exercise above on 10 stocks. If more than half of the flips follow through on most of them, this article is wrong and you should say so.
The honest limits are 3.
First, this article is about default settings on daily charts. A slower acceleration factor, such as a step of 0.01 with a maximum of 0.10, produces far fewer flips and a wider stop. That is a different tool with a different trade-off, and it fixes some of the whipsaw at the cost of giving back more of each trend.
Second, "whipsaw" is defined here by a 3% threshold over 10 days. That number is arbitrary. On a stock that normally moves 4% a day, 3% is nothing. Use the instrument's own ATR instead if you want the test to be fair — article 16 shows how.
Third, a trailing stop that removes you often is not automatically bad. If your average win is 6 times your average loss, a low win rate is fine. Judging any exit rule by how often it is right is the wrong measurement. Judge it by total outcome.
In India
Parabolic SAR is available by default on every Indian retail charting platform, including the charts inside broker apps. It is commonly applied to NIFTY 50 and BANK NIFTY charts and to liquid single stocks.
Two Indian market features change how it behaves.
Price bands. Most Indian stocks outside the derivatives segment have daily price bands of 5%, 10% or 20%. When a stock hits its band, trading effectively stops at that level for the day. The chart records a smaller range than the market actually wanted. A Parabolic SAR calculated on those truncated highs and lows sits closer to price than the real risk warrants. On a stock that has just been moved to a 5% band after an unusual move, the dots become nearly useless.
Gap risk around scheduled events. Indian markets have a small number of dates that reliably produce large gaps: the Union Budget on 1 February, RBI policy days, and general election result days. A trailing stop cannot protect you across a gap. The dot sits at a level, price opens below it, and you exit at the open, not at the dot. Traders who describe SAR as a "stop-loss" often have not tested what happens when the market never trades at their level.
Costs. Every flip is 2 transactions if you reverse. Indian intraday and delivery costs include brokerage, Securities Transaction Tax, exchange charges, GST and stamp duty. On a strategy that flips 25 times a year, costs are not a detail. Calculate them before you judge the returns.
In the United States
The indicator is standard on every US platform and is frequently applied to S&P 500 index products, large-cap stocks and futures.
Three US features matter.
No daily price bands on individual stocks. US stocks have Limit Up-Limit Down bands that pause trading briefly around rapid moves, but there is no equivalent of a 5% daily ceiling for the whole session. A US chart therefore records the full daily range more often, so the true range that feeds any distance-based rule is more honest.
Extended-hours trading. Most US company results are released after the close. A stock can move 10% between 4:15 pm and the next open. Your SAR dot is irrelevant during that window. Nothing is trading at your level, and the regular-session chart will show a gap.
Futures and 23 hour sessions. On S&P 500 futures, the session runs nearly around the clock, which reduces gaps but produces a continuous chart with different noise characteristics from a cash index. A SAR tuned on the cash index will not behave the same way on the future. If you test on one and trade the other, you have tested nothing.
Where they differ, and what that tells you
The real difference is what the chart is allowed to record.
An Indian stock at its upper circuit produces a candle with a small range and a close at the high. Nothing about that candle says "buyers wanted to pay 14% more and were not permitted to." Every distance-based indicator — Parabolic SAR, ATR, Bollinger Bands, Supertrend — reads that small range as low volatility and places its levels close to price. The stock is at its most dangerous and the indicator reports calm.
A US stock in the same situation prints the full move. The indicator widens.
What that tells you is a rule for Indian charts. After any day where a stock closed exactly at a circuit limit, treat every volatility-derived level on your chart as understated for at least the next 5 sessions. The indicator has been fed censored data. It is not broken; it is answering honestly using numbers the exchange truncated.
This is also why blindly copying US indicator settings into an Indian chart quietly changes the risk. The same numbers describe a different amount of reality.
Carry this
- Parabolic SAR is an exit rule wearing the costume of a signal.
- It has no idea whether a trend exists. Ask ADX, or ask the chart.
- Count the flips on your own chart over a year before you trust a single one.