What it measures
The name is usually translated as ‘equilibrium chart at a glance’. It was worked out over many years by the Japanese journalist Goichi Hosoda, who wrote as Ichimoku Sanjin, and published in 1969. It looks crowded, and it is made of very little.
Every line but one comes from the same sum: take the highest high and the lowest low of the last N bars and find the point halfway between them. That midpoint is where the middle of the market’s recent range has been. It is not an average of closes: one bar with an extreme high moves it as much as it can be moved, and then nothing moves it until that bar leaves the window.
The conversion line is the midpoint of 9 bars and the base line the midpoint of 26. Leading span A is halfway between those two, and leading span B is the midpoint of 52 bars. What makes the chart unusual is where the two spans are put: 26 bars to the right of the bar they were calculated at. The fifth line, the lagging span, is the close, put 26 bars to the left.
So the cloud that stands beyond the last bar is not a forecast. It is the midpoints of bars that have already closed, drawn 26 bars later than they were worked out. The machine on this page leaves that space empty of candles, because no prices exist there.
How it is calculated, step by step
- 01
Learn the one sum it uses.
Midpoint of N bars = (highest high of the last N bars + lowest low of the last N bars) ÷ 2. The latest bar is included.
- 02
Conversion line (tenkan-sen).
The midpoint of 9 bars. It is the quickest of the lines.
- 03
Base line (kijun-sen).
The midpoint of 26 bars.
- 04
Leading span A (senkou span A).
(Conversion line + base line) ÷ 2, drawn 26 bars ahead: the value worked out at this bar is placed 26 bars to its right.
- 05
Leading span B (senkou span B).
The midpoint of 52 bars, also drawn 26 bars ahead. It is the slowest line, and it runs flat for long stretches.
- 06
Shade the space between the two spans.
That is the cloud (kumo). It is given one tone where span A is above span B and another where span B is above span A. Where they cross, the cloud narrows to nothing and changes tone.
- 07
Lagging span (chikou span).
The close, drawn 26 bars back: this bar’s close is placed 26 bars to its left. Nothing is calculated.
Programs differ on how the 26 is counted. Many count the latest bar as the first of the 26 and so move the spans 25 bars; this page moves them a full 26. On this page the displacement always equals the length of the base line, which is the usual convention; some programs let it be set separately. The numbers 9, 26 and 52 are commonly explained by the six-day trading week of the time: a week and a half, a month and two months.
A worked example, by hand
Five bars, written high / low / close: 10 / 8 / 9, then 12 / 9 / 11, then 11 / 9 / 10, then 13 / 10 / 12, then 14 / 11 / 13. Small settings so it fits on a page: conversion 2, base 3, span B 4, and a displacement of 3 bars.
- Conversion (2 bars): bar 2 = (12 + 8) ÷ 2 = 10; bar 3 = (12 + 9) ÷ 2 = 10.5; bar 4 = (13 + 9) ÷ 2 = 11; bar 5 = (14 + 10) ÷ 2 = 12
- Base (3 bars): bar 3 = (12 + 8) ÷ 2 = 10; bar 4 = (13 + 9) ÷ 2 = 11; bar 5 = (14 + 9) ÷ 2 = 11.5
- Span A, worked out at bars 3, 4 and 5: (10.5 + 10) ÷ 2 = 10.25; (11 + 11) ÷ 2 = 11; (12 + 11.5) ÷ 2 = 11.75. Drawn 3 bars ahead, at bars 6, 7 and 8
- Span B (4 bars), worked out at bars 4 and 5: (13 + 8) ÷ 2 = 10.5; (14 + 9) ÷ 2 = 11.5. Drawn at bars 7 and 8
- The cloud at bar 7 lies between 10.5 and 11, and at bar 8 between 11.5 and 11.75, with span A on top both times
- Lagging span: the close of bar 4, 12, is drawn at bar 1; the close of bar 5, 13, is drawn at bar 2
There are five bars, and the cloud reaches bar 8. Nothing in it comes from bars 6, 7 or 8, which do not exist: it is the midpoints of bars 1 to 5, drawn three bars to the right.
The numbers in this example were chosen to be easy to add up. They are not prices of anything.
How people read it
- The price against the cloud. A price above the cloud is described as an uptrend, one below it as a downtrend, and one inside it as a market without a clear direction. The cloud under or over today’s bar was calculated 26 bars ago.
- The cloud itself. Its tone shows which span is on top, and so whether the shorter midpoints are above the longer one. A thick cloud means the two have been far apart; a thin one, or a change of tone, means they have met. Readers treat the cloud’s edges as areas where a move may pause.
- The conversion line against the base line. The quicker midpoint crossing the slower one is read like the crossing of two moving averages, and is given more weight when it happens on the side of the cloud that agrees with it.
- The lagging span. Because it is today’s close drawn 26 bars back, seeing it above the candles there says one thing: the latest close is higher than the price was 26 bars ago.
- The cloud ahead. It shows where the midpoints calculated up to now will be drawn. Readers look at its tone and thickness; neither says anything about the bars that will be drawn in front of it.
What it cannot tell you
- It cannot see ahead. The part of the cloud to the right of the last bar is made entirely from bars that have closed. If the price falls sharply tomorrow, that cloud does not move: it was fixed when its bars closed.
- It cannot be early. Span B needs 52 bars, and the cloud beside any bar is 26 bars old. By the time the price is clear of the cloud, much of a move has happened.
- It cannot count as five opinions. All five lines are made from the same highs, lows and closes, and four of them from the same sum. When they agree, that is one fact drawn five ways.
- It cannot tell a pause from a turn. In a sideways market the price passes into and out of the cloud repeatedly, and the conversion and base lines cross again and again.
- It cannot say that 9, 26 and 52 are right. They were chosen for one market’s calendar in another century, and they are kept by convention.
Common mistakes
- Reading the cloud ahead as the place the price is expected to go. It is a record, displaced.
- Testing a rule with the lagging span as if it were known at the bar where it is drawn. The value drawn at a bar is a close from 26 bars later. A test that uses it there is using prices that had not happened.
- Treating agreement among the five lines as confirmation.
- Comparing two charts that displace by 25 and by 26 bars, or that use different settings, as if they showed the same cloud.
- Treating the edge of the cloud as a level the market must respect. It is a midpoint of old highs and lows, and it is flat only because nothing has exceeded them yet.
Questions people ask
- What do 9, 26 and 52 mean in Ichimoku?
- They are three lengths, in bars. The conversion line is the midpoint of the highest high and lowest low of 9 bars, the base line of 26 bars and leading span B of 52 bars. The 26 is also how far the two leading spans are drawn ahead and the lagging span behind. The numbers are the published settings, commonly explained by the six-day trading week of the time, and are kept by convention.
- Does the Ichimoku cloud predict where the price will go?
- No. The cloud drawn ahead of the last bar is calculated only from bars that have already closed: it is the midpoint of the conversion and base lines, and the midpoint of the last 52 bars, drawn 26 bars to the right. It shows where past midpoints will be plotted, not where future prices will be.
- What is the lagging span (chikou span) for?
- It is the latest close drawn 26 bars back, so that it can be compared by eye with the candles of that time. Above them, the close is higher than the price was 26 bars ago; below them, lower. It contains no calculation, and at the place where it is drawn it shows a close that had not yet happened then.
- Why is the cloud two colours?
- The cloud is the space between leading span A and leading span B. It is shaded one tone where span A is the higher and another where span B is. Span A follows the shorter midpoints and span B the 52-bar one, so the tone shows which has been above the other, 26 bars before the place where it is drawn.
The words on this page
An indicator is arithmetic on prices that have already happened. It describes what a price did; it does not predict what a price will do. The chart on this page is invented: a seeded random walk, not a market. This page is an explanation for study. It is not advice, a recommendation or a forecast, and nothing an indicator shows says anything certain about what a price will do next.