The Ichimoku Cloud is a five line system that puts trend direction, momentum, support and resistance on a single chart, with part of it projected into the future. Japanese journalist Goichi Hosoda spent decades refining it and published the finished method in a 1969 book under the name Ichimoku Kinko Hyo, which translates roughly as one glance equilibrium chart. The cloud itself, the shaded area between two of the five lines, is the piece most traders actually read.
Every line is a midpoint of a range, not an average of closes. That is the structural difference from a moving average system and it matters: a midpoint moves only when a new extreme enters or leaves the window, so Ichimoku lines go flat for long stretches and those flat sections often act as real support or resistance.
The five lines
| Line | Japanese name | Formula | Plotted where |
|---|---|---|---|
| Conversion Line | Tenkan Sen | (9 period high + 9 period low) / 2 | On the current bar |
| Base Line | Kijun Sen | (26 period high + 26 period low) / 2 | On the current bar |
| Leading Span A | Senkou Span A | (Conversion Line + Base Line) / 2 | 26 periods into the future |
| Leading Span B | Senkou Span B | (52 period high + 52 period low) / 2 | 26 periods into the future |
| Lagging Span | Chikou Span | Current close | 26 periods into the past |
The highs and lows are the extreme high and extreme low across the whole window, including the current bar. Some implementations exclude the current bar, which produces a slightly different line, so verify before you compare charts.
A worked example
Suppose you are on a daily chart with the following range extremes and a current close of 106.20.
| Input | Highest high | Lowest low | Midpoint | Result |
|---|---|---|---|---|
| Last 9 bars | 108.40 | 101.60 | (108.40 + 101.60) / 2 | Conversion Line 105.00 |
| Last 26 bars | 112.00 | 96.00 | (112.00 + 96.00) / 2 | Base Line 104.00 |
| Last 52 bars | 118.00 | 92.00 | (118.00 + 92.00) / 2 | Leading Span B 105.00 |
| Two lines above | Conversion 105.00 | Base 104.00 | (105.00 + 104.00) / 2 | Leading Span A 104.50 |
Read what that configuration is telling you. Both Leading Spans are plotted 26 bars ahead, where Span A sits at 104.50 and Span B at 105.00. Span A below Span B means the future cloud is bearish, and the two are only 0.50 apart, so it is a thin cloud offering weak resistance. Price at 106.20 is above the current cloud. Conversion above Base is a mild bullish tilt. The honest summary is a market that has just turned up inside a longer range, with a flimsy overhead cloud forming, which is a setup that resolves quickly in either direction.
import pandas as pd
def ichimoku(df, conv=9, base=26, spanb=52, shift=26):
def mid(n):
return (df["high"].rolling(n).max() + df["low"].rolling(n).min()) / 2
out = pd.DataFrame(index=df.index)
out["conversion"] = mid(conv)
out["base"] = mid(base)
out["span_a"] = ((out["conversion"] + out["base"]) / 2).shift(shift)
out["span_b"] = mid(spanb).shift(shift)
out["lagging"] = df["close"].shift(-shift)
return outReading the cloud
The cloud, or kumo, is the shaded region between the two Leading Spans. Four properties carry information.
Position of price. Above the cloud is a bullish regime, below is bearish, inside is neutral and usually not worth trading. This is the primary filter and most disciplined Ichimoku users refuse to take long trades while price is under the cloud.
Color or order. When Span A is above Span B the cloud is conventionally drawn green and is called bullish. When Span B is above Span A it is red and bearish. Because both spans are projected forward, you can see this state 26 bars before price gets there.
Thickness. A thick cloud reflects a wide 52 period range and tends to act as substantial support or resistance. A thin cloud gets sliced through. Thickness is the closest thing Ichimoku has to a volatility measure, and it plays a similar role to band width in Keltner Channels.
The twist. Where the two spans cross in the projected region, the cloud changes color. That crossing point is fixed in advance, so traders watch for price to arrive at a twist as a moment where the trend structure is weakest.
The four standard signals
| Signal | Trigger | Strength convention |
|---|---|---|
| TK cross | Conversion Line crosses the Base Line | Strong above the cloud, weak below it |
| Cloud breakout | Price closes outside the kumo | Stronger through a thick cloud than a thin one |
| Kumo twist | Leading Spans cross in the projection | Advance warning, not an entry |
| Chikou confirmation | Lagging Span clears the price bars from 26 back | Used as a veto on the other three |
The chikou check is the one beginners skip and the one experienced users say matters most. If the lagging span is buried inside old price action, the current move is happening inside a zone where a lot of positions were established, and it faces overhead supply. Requiring clear air behind the lagging span filters out a large share of marginal entries.
The settings debate
The 9, 26 and 52 defaults come from the Japanese equity market of Hosoda’s era, when Saturday sessions meant a six day trading week. Twenty six was roughly a month of sessions and 52 roughly two. Those calendar reasons no longer apply anywhere, which is why the settings get argued about constantly.
Three camps exist. Traditionalists leave the defaults alone on the grounds that the system was validated as a whole and that widely watched levels acquire their own significance. A second camp shifts to 7, 22 and 44 to reflect a five day week. A third, common in crypto and futures where the market runs continuously, doubles everything to 20, 60 and 120.
There is no published basis for preferring any of them. The one defensible discipline is to pick a setting per market before you test, keep it fixed across the whole sample, and never tune it after seeing the results.
What the evidence supports
Honest assessment: peer reviewed evidence for Ichimoku specifically is thin, and the studies that exist are typically single market, single period tests that do not survive a serious data snooping correction.
The broader literature is more informative. Brock, Lakonishok and LeBaron reported significant results for simple moving average and range breakout rules on the Dow in the Journal of Finance in 1992. Sullivan, Timmermann and White then re examined that same universe with a bootstrap built to correct for data snooping and found the apparent edge did not carry forward. Park and Irwin’s survey in the Journal of Economic Surveys counted 95 modern studies with 56 positive, 20 negative and 19 mixed, while stressing that selective reporting inflates the positive share. Since Ichimoku is essentially a bundle of range midpoint rules, it inherits those caveats rather than escaping them.
Use the cloud as a structured way to describe trend state. That is a real service, and the projection makes it a genuinely different lens from MACD or an oscillator. Do not treat any of the four signals as a validated edge.
Common problems and fixes
| Problem | Cause | Fix |
|---|---|---|
| Cloud is drawn at a different offset | Some platforms shift by 25 or by the base period setting | Set the displacement explicitly to 26 rather than trusting the default. |
| Backtest results look impossible | Lagging span leaked future data | Delay every chikou based condition by 26 bars. |
| Endless whipsaws | Trading TK crosses while price sits inside the cloud | Require price outside the kumo before any entry. |
| Chart is unreadable | Five lines plus candles plus other indicators | Hide the Conversion and Base lines until you need them, keep the cloud. |
| Lines flat for weeks | No new extreme entered the window | That is correct behavior. Flat lines are meaningful levels, not a bug. |
Frequently asked questions
Is the Ichimoku Cloud a leading indicator?
Only in the narrow sense that the two Leading Spans are drawn 26 bars into the future. Their values are computed entirely from past prices, so nothing is being forecast. What you gain is knowing in advance where the cloud will sit, which is useful for planning levels rather than predicting direction.
What time frame works best for Ichimoku?
Daily and four hour charts are the most common because the 52 period lookback needs a meaningful history. On one minute charts the ranges are dominated by microstructure noise. Whatever you pick, keep the same settings across the whole test rather than optimizing per time frame.
Can I use Ichimoku alone?
Some traders do, since the system covers trend, momentum and levels. In practice most people add a volume or flow measure, because Ichimoku is entirely price based and cannot see participation. Pairing it with the Money Flow Index covers that gap.
What does a thick cloud mean?
It means the 52 period range was wide when that section was calculated, so a lot of trading happened across those prices. Thick clouds tend to slow or reject price. Thin clouds get cut through easily. Thickness says nothing about direction on its own.
Does Ichimoku work on stocks as well as forex?
The calculation is market agnostic and works anywhere you have high, low and close data. It became popular in forex because those markets trend cleanly for long stretches, but nothing in the math is currency specific. Gaps in equities make the cloud less continuous than it looks on a currency chart.
The bottom line
Ichimoku packs a trend filter, a momentum cross, a projected support zone and a confirmation check into one overlay, and once you stop being intimidated by the five lines it is genuinely economical. The cloud alone answers the question most charts leave vague: is this market in a trend, and where does the structure sit ahead of price.
The trap is the density of the signals. Four triggers on one chart invites you to find a reason to trade at any moment. Set the cloud position as a hard filter, use the chikou check as a veto, and read candlestick structure for the bar level detail the system deliberately smooths away.
