Hull Moving Average vs Ichimoku
Michael Whitman
·
Trading since 2017, full-time since 2020
·
YouTube creator, 137K subscribers
·
Published 2026-09-08
·
Updated
·
Reviewed
The hull moving average is one line built from weighted averages so that lag comes out and the line turns early. Ichimoku is a five-part framework whose cloud, plotted 26 sessions ahead, gives a slower read: a close beyond its far side means price has crossed the whole band.
One of these is among the fastest smoothed lines in common use. The other is a framework whose main feature, the cloud, is deliberately slow. Comparing them is comparing two answers to the same question, has the swing turned, and the useful measure is how late each answer arrives and how often it is given when nothing has turned at all.
What each one is
The hull moving average is one line. Alan Hull’s own page describes building it in 2005 to cut lag: take twice a weighted average over half the period, subtract a weighted average over the full period, then smooth the result with a weighted average over the square root of the period. The hull moving average covers the construction.
Ichimoku draws five things, and only the cloud matters here. Its edges are span A, the average of the 9- and 26-session midpoints, and span B, the 52-session midpoint, both drawn 26 sessions to the right of the bar they were computed on. Ichimoku covers the other three lines, and the kumo cloud covers the band itself.
They are not built from the same measurement. The hull is a weighting of closes. The cloud’s boundaries are midpoints of highs and lows over 26 and 52 sessions, computed 26 sessions earlier.
Where they differ
Speed. Differencing two averages pulls the hull forward, so its slope turns within days of a price turn. The cloud is drawn from data that is already 26 sessions old.
Whether the tool can say nothing. The hull is always sloping one way. Price inside the cloud is an explicit no-trend reading.
Level against zone. The hull is one price. The cloud has a top and a bottom, so crossing it is a process: a close past the far side means price has traveled the whole width.
Whether a value can sit outside traded prices. The hull can overshoot, because its differencing step produces values no recent close supports. Every cloud boundary is a midpoint of prices that traded.
Where the fast line and the cloud agree
Neither uses a bar before it has finished. The cloud’s forward shift draws old information ahead; it does not forecast. Both whipsaw in a range: the hull by turning, the cloud by letting price drift through it.
Which one to use: Hull Moving Average or Ichimoku?
Run the hull when an early turn matters more than a false one. Above your own structure work, a fast line turning is a useful trigger, as long as something else has decided which turns count.
Run Ichimoku when you need the tool to filter conditions. If your problem is taking trades in chop, a framework that can say price is inside the cloud does a job no single line can.
Run both when you want speed with a check. The hull’s turn can be the alert and the cloud the confirmation, as long as you accept that the confirmation came later than the alert in four swings in five on SPY.
And when deciding which to learn first, take Ichimoku. The hull teaches lag; Ichimoku teaches conditions, and conditions decide whether the lag mattered.
Speed against the number of turns
A faster line turns more often in both directions. The hull’s early turn at a real swing comes with many turns at moves that never became swings.
The cloud trades that the other way. It reacts late, but a close beyond its far side requires price to cross a band built from five to ten weeks of ranges, which small wiggles rarely do.
A worked example: early 2026
A top. SPY closed at $695.49 on 27 January 2026, the peak close before its first fall of 7% or more since April 2025. The 26-session hull’s slope turned down on 4 February, six sessions later, with SPY 1.34% below that close. The first close below the bottom of the cloud came on 6 March, 27 sessions after the high, 3.32% below it.
And a bottom. SPY’s lowest close of that fall was $631.97 on 30 March. The hull turned up on 6 April, four sessions later and 4.27% higher. The first close above the top of the cloud was 13 April, nine sessions later, 8.57% above the low.
Both turns are known only in hindsight. Nobody knew on 27 January that it was the top; the measurement starts from the extreme because that is the only fixed point a lag can be measured from.
The original data
The sample. SPY daily closes, highs and lows from Yahoo Finance, January 1993 to 2 October 2026; turns are found from 3 January 1994. Hull: 26-session HMA per Alan Hull’s formula, so WMA(13) and WMA(26), smoothed over 5 sessions (the integer square root of 26). Cloud under each session: the span A and span B values from 26 sessions before it, per StockCharts.
The swings. A turn is a close-to-close extreme followed by a move of at least 7% the other way before a new extreme: 124 turns, 62 peaks and 62 troughs. They are picked with hindsight. For each, the hull signal is the first session after the extreme on which the hull’s slope pointed the new way; the cloud signal is the first close beyond the cloud’s far side (below its bottom after a peak, above its top after a trough).
How late each came. Hull: a median of 5 sessions after the extreme, at most 20, with price a median 3.70% away from the extreme. Cloud: a median of 15 sessions, at most 185, a median 6.31% away. The hull came first at 100 turns, the cloud at 19, and they fell on the same session at 5.
When a signal missed the swing. Of the 123 turns followed by another, the cloud signal had not come by the time the next turn printed on 19; the hull’s had not on 6.
The cost of the hull’s speed. Across the same span its slope changed direction 719 times, 21.96 a year, while SPY made 124 swings of 7% or more, under four a year. Most hull turns were not at a swing of that size. Every turn and both signal dates are in the SPY hull and cloud file.
When it fails
The characteristic failure is acting on every hull turn. The line turned about 22 times a year and SPY swung 7% under four times a year, so most turns acted on are wiggles, each paying a round trip.
A second failure is waiting for the cloud at the top of a fast fall. In this sample the far-side close came a median 6.31% from the extreme, and in 19 swings it came only after the next turn.
A third is porting both defaults to intraday charts unexamined. The 26-session lengths were set for daily bars.
A fourth is stopping at the hull line, which can overshoot the price range entirely.
And a fifth is judging either on a trending sample. Both look excellent in one sustained move.
Related
The hull moving average covers the reduced-lag construction and the overshoot it buys. Ichimoku covers each component, including the state where price sits inside the cloud. And the kumo cloud covers the band whose far side is the slower signal here.
Count how often your fast line turns, not only how early. A line that turns 22 times a year will catch every real swing early and many swings that never happen. Decide in advance which turns you act on, and let a slower reading like the cloud filter the rest.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.