Ichimoku Cloud — a whole system on one chart
The answer
Ichimoku is 5 lines drawn from the midpoints of past price ranges, with 3 of them shifted forward or backward in time. The shaded area between 2 of those lines is the cloud, and it marks where price found balance over the last few months.
It looks like a forecast because part of it is drawn to the right of today's price. It is not a forecast. Every value in it was computed from prices that have already happened.
Why this costs you money
Ichimoku is the most visually persuasive thing you can put on a chart, and visual persuasion is not evidence.
The specific loss works like this. The cloud extends 26 bars into the future. A trader sees a thick cloud ahead, above current price, and concludes the market will meet resistance there next month. They avoid a position, or they sell into strength.
The cloud drawn over next month was calculated from the highs and lows of the last 52 days. It contains no information about next month at all. It is a picture of the past, moved to the right. Trading it as a forecast is trading a delayed echo and calling it a signal.
The second loss is quieter. A beginner adds all 5 lines, sees a chart covered in overlapping colour, and cannot make a decision. They wait for all 5 conditions to line up. On most charts, that alignment appears long after the move has started, because 3 of the 5 conditions are lagged by design. The trader enters late, near the end of the trend, with a stop that must be very wide because the Kijun-sen is far away.
Late entry plus wide stop is the worst combination in trading. It is not a failure of the system. It is what a lagging system does when you demand certainty from it.
How it works
The full name is Ichimoku Kinko Hyo, which translates roughly as "one glance equilibrium chart". It is credited to Goichi Hosoda, a Japanese newspaper writer who worked on it with a team of assistants over many years and published it in 1969.
Every line is built from the same simple operation: take the highest high and the lowest low over a period, add them, divide by 2. That is a midpoint of a range, not an average of closes. This is the one thing that makes Ichimoku different from a moving average system, and it is worth holding on to.
The 5 lines:
| Line | What it is | Where it is drawn |
|---|---|---|
| Tenkan-sen (conversion) | Midpoint of the last 9 bars' range | On today's bar |
| Kijun-sen (base) | Midpoint of the last 26 bars' range | On today's bar |
| Senkou Span A | Average of Tenkan and Kijun | Shifted 26 bars forward |
| Senkou Span B | Midpoint of the last 52 bars' range | Shifted 26 bars forward |
| Chikou Span (lagging) | Today's closing price | Shifted 26 bars backward |
The cloud, or Kumo, is the shaded area between Senkou Span A and Senkou Span B. When Span A is above Span B the cloud is usually coloured one way; when Span B is above, the other way.
The numbers 9, 26 and 52 come from a Japanese trading week that included Saturdays, so a month was roughly 26 trading days. They are conventions. Nothing breaks if you change them, and nothing improves reliably either.
Two things about the shifts deserve a plain sentence each.
The cloud ahead of price is old data. Senkou Span B plotted over next Tuesday was calculated from the 52 bars ending today. It moves forward with time. It cannot know anything.
The Chikou Span is not new information. It is today's close, drawn 26 bars to the left. Asking "is the Chikou Span above the price it is sitting on?" is exactly the same as asking "is today's close above the close of 26 bars ago?" That is a momentum question you can answer without drawing anything. Presenting it as a separate line makes 1 fact look like 2 pieces of evidence.
What it tells you, and what it does not
It tells you where price has found balance recently, using ranges rather than closes. Range midpoints have a real advantage: they are less affected by a single strange close, and they capture where trading actually occurred.
It tells you, at a glance, whether price is above, below or inside that balance area. Above the cloud is a trending-up condition. Below is trending-down. Inside is the condition where nearly every trend rule fails, and Ichimoku is unusually honest about naming it.
It tells you the thickness of the recent balance zone. A thick cloud means the 52 bar range and the shorter ranges disagreed a lot, which means price covered a lot of ground. A thin cloud means the ranges converged, which usually follows a quiet period.
It does not forecast. Repeat this until it is automatic.
It does not give you 5 independent opinions. The Tenkan, Kijun and both Senkou spans are all range midpoints over overlapping windows of the same price series. When they agree, that is arithmetic, not confirmation. Article 18 in this cluster explains why this matters more than any single indicator's accuracy.
It does not work well on a chart with fewer than about 100 bars of history, because Senkou Span B needs 52 bars and is then shifted 26 more.
The decision rule
Use the cloud as a filter, and only 1 of the 5 lines as a trigger.
The filter. If price is above the cloud, take only long setups. If below, only short setups. If inside the cloud, take nothing. That third instruction is the most valuable line in the whole system and the one most often ignored.
The trigger. Use the Kijun-sen. Price crossing back above the Kijun in an above-cloud market is a re-entry. Price closing below the Kijun is an exit.
Ignore the Chikou Span as evidence. It is a restatement of today's close. If you want the information it carries, ask whether today's close is above the close of 26 bars ago, and count it as 1 fact, not 2.
Start with the cloud alone for a month. Add the Kijun after that. Most people never need the other 3 lines, and every line you add makes the chart harder to act on.
Try this now
Five minutes on a chart you care about.
- Open a daily chart, 2 years of data. Add Ichimoku with default settings (9, 26, 52, 26). If your app lets you, turn off the Chikou Span and the Tenkan-sen for now. Leave the cloud and the Kijun-sen.
- Shade or note every stretch where price was inside the cloud. On most charts this covers 25% to 40% of the 2 years.
- Measure what price did during those stretches. Take the first and last close of each stretch and note the change.
- Now do the same for the stretches where price was clearly above the cloud, and clearly below it.
- One more count. Find every time price crossed from inside the cloud to above it. Count how many held above for at least 10 trading days.
What you should see. The inside-cloud stretches usually go nowhere. First close to last close is often within 2% or 3%, after a lot of movement in between. That is the picture of a market that would have taken money from any trend rule.
The above-cloud and below-cloud stretches usually contain most of the year's directional movement. And the cross count in step 5 will show you the cost of the filter: a meaningful number of cloud exits fail and fall straight back in.
The lesson is not "Ichimoku predicts". It is that the cloud is a usable description of when not to apply a trend rule, and that description is worth more than most signals.
Three real cases
1. Goichi Hosoda, published 1969 — the source, and what got lost Hosoda designed a complete system with a strict internal logic, in which the 5 elements were meant to be read together and the cloud was a component, not the product. What travelled around the world is the cloud, because it is the part that renders beautifully on a screen. The gap between "a system with rules about when it does not apply" and "a pretty shaded area" is where most retail use of Ichimoku sits.
2. Nikkei 225, 1989 to February 2024 — the home market, and a 34 year lesson in lag The Nikkei 225 peaked near 38,915 at the end of 1989 and did not exceed that level again until February 2024. Across that period the index produced several multi-year advances and several multi-year declines. Ichimoku, on a weekly or monthly chart, would have identified the sustained legs and would also have spent long stretches with price inside the cloud, giving no signal at all. That is the honest profile of the system: excellent at describing established direction, silent during the long middles, and always late at the turns.
3. NIFTY 50 and S&P 500, February and March 2020 — when the cloud is irrelevant Both indices fell extremely quickly in late February and March 2020. Price gapped below the cloud in a matter of days. The cloud below price, which had been described for weeks as support, was crossed without any pause at all. This is not a criticism of Ichimoku. It is the general rule for every support level derived from past prices: a level holds when normal supply and demand meet there, and it is irrelevant when the reason for the move is larger than the level. Any indicator drawn from history is silent about an event that has no history.
The question that resolves it
A novice looks at the cloud stretching to the right and asks: what is the market going to do when it gets there?
An expert looks at the same cloud and asks: what were the highs and lows over the last 52 bars, and is price above them or inside them now?
The second question is the only one the drawing can answer. Once you have asked it that way a few times, the forward projection stops looking like a prediction and starts looking like what it is: a convenient way of showing you where the recent balance zone sits, positioned so it does not overlap today's candles.
What would make this wrong
If Ichimoku carried information beyond price, then an above-cloud condition would be followed by higher prices more often than the base rate for that instrument, measured over many stocks and many years, after allowing for the fact that above-cloud is itself a description of a market that has already risen.
That last clause is where most Ichimoku "evidence" falls apart. Of course above-cloud periods contain gains. Being above the cloud is nearly the same statement as "price has risen recently". A fair test needs a comparison against a simple rule using the same information, such as "price above its 50 day moving average". If Ichimoku does not beat that, it has added complexity and nothing else.
The honest limits are 3.
First, the system was designed for daily and weekly charts of liquid instruments. On a 5 minute chart the shifts still function arithmetically and the concept of a monthly equilibrium does not.
Second, the settings 9, 26 and 52 are conventions from a market with a 6 day week. There is no evidence they are optimal, and changing them to 10, 30 and 60 produces a chart that behaves almost identically. If a small change to the settings destroyed the results, the results were noise. If it does not change them, the specific numbers were never the point.
Third, the visual density of Ichimoku creates false confidence. A chart with 5 lines and a shaded band feels more analysed than a chart with 1 line. It is not. It is the same price series, described 5 ways.
In India
Ichimoku is available on all major Indian broker charting platforms and free charting sites. It is used on NIFTY 50, BANK NIFTY and liquid single stocks, though it is less popular in India than Supertrend.
Three Indian features matter.
Range midpoints and price bands. Every Ichimoku line is a midpoint of highs and lows. When a stock is limited by a 5%, 10% or 20% daily band, the recorded high or low is administrative. The cloud built from a series of circuit-limited days is a picture of what the exchange permitted, not of where the market found balance.
Holidays and the shift count. The 26 bar shift counts trading bars, not calendar days. The Indian market has a larger number of scattered public holidays than the US market, so a 26 bar shift covers a longer calendar span in India, irregularly. This does not break the indicator, but it does mean that "the cloud is projected a month ahead" is approximate here in a way it is not elsewhere. Concentration in the index. NIFTY 50 is heavily weighted towards a small number of very large companies. A cloud drawn on the index can describe balance in 5 stocks and describe nothing about the other 45. This applies to every index-based indicator and is covered in article 19.
In the United States
Ichimoku is standard on US platforms and is applied to equities, index futures and currencies.
Three US features matter.
Overnight gaps. Most US companies report after the close. A gap moves price from one side of the cloud to the other without ever trading at the cloud. Any rule that says "enter when price crosses the cloud" needs a defined behaviour for this case, because the cross happens at a price you could not transact at.
Nearly continuous futures sessions. On S&P 500 futures the session covers most of 24 hours, so highs and lows include overnight moves. An Ichimoku on futures and one on the cash index will produce different clouds from the same underlying market. Choose the one you actually trade.
The instrument was designed for a Japanese equity market of the 1950s and 1960s. Applying it to US index futures is not wrong, but it is a transplant. Nothing in the design accounts for continuous electronic markets, after-hours earnings or index options expiries.
Where they differ, and what that tells you
Both markets can move a stock from above the cloud to below it without trading in between. They get there by opposite routes, and the routes tell you different things.
In the United States, that happens through an overnight gap on company news. The information was released at 4:05 pm and the price adjusted before anyone could act. The cloud was not defended because no trading occurred at it.
In India, it happens through a circuit limit. The stock does not gap past the cloud in one jump. It is held at a band for 1 or more sessions and then continues. On the chart you see a stair, not a gap. And crucially, the days spent at the band feed narrow ranges into every Ichimoku line, so the cloud drawn 26 bars later is built from a period when the market was not permitted to express itself.
What that tells you is 2 different corrections. On a US chart, discount cloud levels formed around an earnings gap, because 1 abnormal bar has stretched them. On an Indian chart, discount cloud levels formed around circuit days, because the ranges that made them were censored downward.
Same indicator, same formula, 2 different reasons the level under it is not real.
Carry this
- Every Ichimoku line is a midpoint of a past range. Nothing in it looks forward.
- Price inside the cloud is the system telling you not to trade. That is its most valuable output.
- Five lines from one price series are 1 opinion, not 5 confirmations.