Ichimoku Cloud

Jane switches on Ichimoku and initially sees a tangle of lines. She starts with two: Tenkan-sen and Kijun-sen. Fidelity gives the usual windows as 9 and 26 periods. TradingView explains that the highest High and lowest Low of the whole window are added and divided by 2. The result lies halfway between those extreme prices. It is not SMA, the simple average of all closing prices, or Close values. If prices inside the range change while its extremes stay the same, this midpoint remains unchanged. That explains why a line can run flat even while trading continues. [1] [2]

Once Jane understands the basic lines, she can add the cloud. According to Fidelity, Senkou Span A averages the current Tenkan-sen and Kijun-sen values. Senkou Span B instead takes the midpoint between the highest and lowest prices of a longer window, commonly 52 periods. Colouring the area between them creates the “cloud”. Its thickness is the difference between calculated price levels. It is not a statistical confidence interval or a band with a promised probability of containing future prices. Colour makes the chart easier to read; it does not add information to the calculation. [1]

TradingView normally plots Senkou Span A and Senkou Span B 26 periods ahead. Imagine Jane calculating today’s value on paper and sticking it into a later box on the timeline. She has moved the result, not discovered the future. A cloud can therefore appear to the right of the latest price candle even though no future trades have taken place there. Keep the calculation time separate from the plotting time. A green area on the right is neither an observed future price nor a promise of a rise. [2]

Another line, Chikou Span, moves in the opposite direction. Fidelity describes it as the current Close, the candle’s closing price, plotted 26 candles back. Jane therefore sees a newer value at an older chart position. On a completed picture it can look as though it had already warned of the movement that followed. A trader at that earlier time could not have known it. Historical strategy tests must use only information available when the decision was made. Shifting a drawing is not itself the mistake; treating it as earlier knowledge is. [1] [4]

TradingView reads price above the cloud as an upward context, below it as downward, and inside it as transitional. Jane must nevertheless draw an imaginary vertical line through the same moment. She compares today’s price with the cloud displayed at today’s candle, not with its newly calculated end on the right. Otherwise she is comparing different time positions. This helps organise a view of the trend, but does not by itself specify an entry price, an acceptable loss or whether the next trade will succeed. [2]

The numbers 9, 26 and 52 given by Fidelity are common settings. A “period” takes the length of the chart’s candle: hourly intervals on an hourly chart, daily ones on a daily chart. The same numerical settings therefore cover different lengths of history on different charts. Jane records the timeframe, window lengths and displacement. Changing them until a historical chart looks ideal does not establish that the settings will work elsewhere. A new combination should also be checked on data that did not guide its selection. [1] [4]

TradingView notes that a live candle’s High, Low and Close can change until it closes. A new extreme can move a range midpoint; the latest price can also change the line drawn backwards. Jane may see a situation during the candle that does not remain on the final historical chart. That is not automatically a tool failure. She must decide whether she acts during the candle or after it closes, and use the same approach when evaluating history. A finished picture does not show every temporary state that occurred during trading. [3]

Fidelity interprets Tenkan-sen and Kijun-sen crossovers in the context of the cloud. Several agreeing lines still come from price history, rather than independent guarantees. Jane needs entry, exit and position-size rules, as well as a condition that invalidates her scenario. Results must include fees and slippage, the difference between expected and actual execution prices. She must exclude lookahead bias, using information unavailable at the time, and overfitting, tailoring rules too closely to one history. Ichimoku can organise the questions; adding lines does not remove lag or false signals. [1] [4]