Ichimoku vs Parabolic SAR
Michael Whitman
·
Trading since 2017, full-time since 2020
·
Published 2026-09-08
·
Updated
·
Reviewed
Ichimoku is a five-part framework that can show a market with no trend, with price inside its cloud. The parabolic SAR is J. Welles Wilder's trailing stop that accelerates toward price and reverses to the other side when touched, so it is always either long or short.
One of these is a framework that can report uncertainty. The other is a mechanism that cannot, because it is always in a position. The usual advice joins them: let the cloud say when a trend exists and let the dots manage the trade. More than three decades of SPY show what that pairing does and does not do.
What each one is
Ichimoku describes the market in five lines. The one that matters against the dots is the cloud: the band between span A (tenkan and kijun averaged) and span B (the middle of the last 52 sessions' range), each drawn 26 sessions ahead, so price can sit above it, below it or inside it. Ichimoku covers the other components.
The parabolic SAR prints a trailing level that accelerates toward price. Wilder introduced it in his 1978 book. Each session the level moves a fraction of the way toward the trend’s extreme point; the fraction starts at 0.02 and rises by 0.02 at each new extreme, to a maximum of 0.20. When price trades through the level, it flips to the other side. Parabolic SAR covers it.
One is a reading and the other is a position. The dots exist to manage a trade; the framework exists to describe the market.
Where they differ
Whether uncertainty can be shown. Ichimoku can show price inside a flat cloud with its lines tangled. The dots always point somewhere.
What sets the level. The SAR’s level depends on time and on the trend’s extreme, and it keeps closing in on price even if price stalls. The cloud’s levels are midpoints of old ranges.
How each behaves at the end of a move. The dots tighten by design and exit early. The framework reports that its components have stopped agreeing.
What the parameters are. Five periods designed as a set against a step and a maximum that Wilder chose as constants.
Where the dots and the cloud agree
Both lag: every value is built from bars already printed, and the cloud’s forward plot is old data drawn ahead. Both struggle in a range, and both set their levels by formula rather than at the point where a trade idea is wrong.
Which one to use: Ichimoku or Parabolic SAR?
Use Ichimoku to decide direction. It is the only one of the two that can say nothing is happening.
Use the dots as an exit inside that direction. As a trailing stop in a method that has already chosen a side, the acceleration gives back less at the end of a run.
Use the dots when your problem is holding too long. The tightening is mechanical and exits before you would have.
And never take the reverse automatically. A stop being touched is not an argument for the opposite position; the framework decides whether there is one.
Does the cloud calm the dots?
The common claim is that the SAR flips constantly when price is inside the cloud. On SPY it did not flip more often there. It reversed at roughly the same rate wherever price sat, because what drives a reversal is price crossing a level that is always closing in, not the absence of a trend.
So the cloud cannot be used to predict fewer whipsaws. What it can do is say which reversals agree with the side you have chosen, and on SPY nearly half did not.
A worked example: April 2026
On 1 April 2026 the dots turned long below the cloud. SPY had been in a falling SAR since 3 March. On 1 April its high of $658.52 crossed the falling SAR, so the indicator reversed: the new SAR was the down-move’s extreme low, $629.28. SPY closed at $655.24, still below the cloud, whose bottom was $684.52 and top $686.64 that day.
A cloud reader would have ignored that reversal. Price below the cloud says the trend is down, and a long signal against it is exactly the kind the pairing is meant to filter.
This time the dots held long until 18 May. One example settles nothing; it shows that a reversal against the cloud is a different reading, not a wrong one, and that the pairing has to decide in advance which of the two is in charge.
The original data
Inputs. SPY daily highs, lows and closes from Yahoo Finance, January 1993 to 2 October 2026; reversals are counted from 3 January 1994. SAR per StockCharts’ description of Wilder’s rules: step 0.02, maximum 0.20, the level never above the prior two lows in an uptrend or below the prior two highs in a downtrend, a reversal when the low (or high) crosses it, the new level set to the prior extreme point. Cloud state: the close against span A and span B as computed 26 sessions before; above both is above the cloud, below both is below it, anything else inside.
How often it reversed. 801 reversals, 24.46 a year.
Where price was when it reversed. Above the cloud: 547 (278 to long, 269 to short). Inside: 76 (30 to long, 46 to short). Below: 178 (93 to long, 85 to short). As shares of all reversals that is 68.29%, 9.49% and 22.22%; as shares of all sessions, price sat above the cloud 66.26% of the time, inside it 10.65% and below it 23.09%.
| Close against the cloud | Reversals | Per 100 sessions in that state |
|---|---|---|
| Above | 547 | 10.01 |
| Inside | 76 | 8.66 |
| Below | 178 | 9.35 |
Against the cloud. 362 reversals, 45.19%, turned long with price below the cloud or short with price above it. Counted by session, the dots were long while price was below the cloud on 785 sessions and short while it was above on 1,674.
How long a reversal lasted. The median time to the next reversal was 8 sessions after one above the cloud, 11 after one inside and 10 after one below. Every reversal, the cloud that day and the gap to the next one are in the SPY SAR reversals file.
When it fails
The failure is trading every reversal as a trend signal. The dots reversed about every two weeks on SPY, wherever price sat against the cloud, and each reversal acted on costs a round trip.
The second failure is assuming the cloud will quiet the dots. It did not; it only labels which reversals agree with it.
A third is taking the reverse automatically. That is a second decision.
A fourth is default acceleration on every instrument. The 0.02 and 0.20 are Wilder’s constants, not measurements of your market.
And a fifth is reading the two on different timeframes. They will disagree by design, because the dots on an hourly chart reverse many times inside one daily cloud reading, and every one of those reversals looks like a contradiction that is really just a change of scale.
Related
Ichimoku covers the five-component framework. Parabolic SAR covers the accelerating stop. And the ATR trailing stop covers a trailing distance you set from the instrument’s own range.
Use the dots as an exit inside a direction you have already chosen, not as the thing that chooses it. They reverse every couple of weeks whatever the cloud says, so a reversal on its own carries no information about the trend.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.