An Ichimoku cloud that adapts to volatility, with every setup scored and its plan drawn.

A signal is decided on a closed bar: once a plate or a marker has printed it stays where it is, and every alert fires once for a closed bar. It appears once its bar is complete, which in practice is when the next bar opens. What moves is what is meant to move: the lines, the cloud and the dashboard on the bar that is still forming, and the entry, stop and target lines, which always belong to the newest signal.

In Adaptive mode the Tenkan, Kijun and Span B periods follow volatility. The indicator places ATR(14) between its lowest and highest value over the last 50 bars; at the volatile end the periods shorten by up to 40 percent (5, 16 and 31 instead of 9, 26 and 52), at the quiet end they lengthen by as much (13, 36 and 73), and in between they move smoothly. Classic mode keeps the three periods fixed. The Volatility row of the dashboard shows where ATR sits.

Nine factors, each worth a fixed number of points: Price against the cloud: 20 above the cloud, 8 inside it, 0 below for a buy; Tenkan against Kijun: 15 when Tenkan is above Kijun for a buy; Chikou agreement: 15 when the Chikou span is above the Source one displacement earlier for a buy; Cloud direction: 10 when Span A is rising faster than Span B over 3 bars for a buy; Cloud thickness: 8 when the cloud is not thin, 2 when it is for a buy; Volume: 10 when volume is above 1.2 times its 20-bar average for a buy; Momentum: 12 above +20, 6 above 0 for a buy; Momentum acceleration: 10 when momentum is positive and rising for a buy; RSI (optional): 10 when RSI(14) is above 50, only with the RSI Confluence Filter on for a buy. The thickness and volume factors do not depend on direction, so they lift or lower the buy and sell scores together. The total is out of 100, or 110 with the RSI factor on, and is graded Strong at 80 and Standard at 50. It counts how much agrees at that bar; it is not a forecast of what happens next.

A TK cross is Tenkan crossing Kijun, with the cloud not thin, volume above 1.2 times its average, the Chikou agreeing and the score at Min Confluence Score; it prints the Long or Short plate. A Kijun bounce (KB) is a wick to the Kijun that closes back on the trend side, above the cloud with Tenkan above Kijun for a buy. A cloud-edge bounce (CB) is the same against the cloud top or bottom. A Kumo breakout (BK Up or BK Dn) is a close through the cloud with volume above 1.3 times its average, and a retest (a small circle) is a return to the broken edge 2 to 10 bars later. A Chikou breakout is available too, off by default.

It is the lowest score a TK cross, a Kijun bounce, a cloud-edge bounce or a Chikou breakout needs. Kumo breakouts and retests do not use it. On the 16 MetaTrader 5 charts we measured, the score-gated signals per 1,000 bars came to 41 to 87 at 40, 41 to 87 at 50, 39 to 81 at 60 and 26 to 54 at 80. Start at the default 40 to see every setup worth a look, then raise it once you know which scores you want to act on.

The signal bar's close is the entry. SL Mode decides the stop: Kumo Edge puts it beyond the far side of the cloud plus the ATR buffer; Kijun puts it beyond the Kijun plus the ATR buffer; ATR puts it two ATRs from the entry. The buffer is 0.5 ATR by default, and the risk is never taken as less than half an ATR. The targets sit at 1.5, 2.5 and 4.0 times the risk. Targets and stop are checked on closed bars; the stop or the third target ends the tracked trade, and every new signal starts a new one with its own levels.

Support and resistance zones. When the Kijun has stayed flat for four bars and then moves, the flat level is marked with a box 0.15 ATR either side of it, from 5 bars before to 10 bars after. A flat Kijun is where the market spent time, which is why these levels are worth watching. Show S/R Zones turns them off.

The classic Ichimoku on the last closed bar of M15, H1 and H4 (you can choose others): Bull when that bar closed above its cloud with Tenkan above Kijun, Bear in the mirror case, Neutral otherwise. On the bar that is still forming they keep the last closed reading, so they do not flicker. When all three agree, the alignment alert fires.

A count, not a promise. It shows first targets reached against first targets plus stops, for every signal the indicator drew on this chart since its first calculated bar, with no spread, commission or slippage. A trade that touches its first target and then its stop counts on both sides. It describes the drawing on that history, and we do not quote it anywhere on this page.

Yes. Each one was saved by MetaTrader itself on 14 October 2026 (MetaTrader 5) and 15 October 2026 (MetaTrader 4), with the shipped indicator attached at its default inputs, and none was edited. The MetaTrader 5 charts show the history stored in that terminal, which ends between 31 Dec 2025 and 8 Sep 2026 depending on the symbol; the MetaTrader 4 charts were live. The dashboard values and signal counts printed next to them were computed from the same charts' bars at the same moment, by the same engine.

No, on purpose. Any single number depends on the market, the timeframe, the period, the settings and on how you manage the trade, so it would flatter some charts and hide others. The frequency tables tell you how often it fires and the charts show signals that reached their targets and signals that hit the stop. Test it on a demo account on your own markets before you trade it.

Every value the engine computes was compared with the original design's own output, bar by bar, on both builds: 15,026,434 values per platform, with no differences in anything the indicator shows or alerts, and the plates, markers, levels and dashboard text were compared the same way. On the charts on this page, the signal counts each chart drew were also recomputed independently and matched on all 26 charts.

Yes. It reads price and volume only, so it runs on any symbol your broker quotes. We captured it on 13 markets across forex, gold, stock indices and crypto, from M15 to H4. On most MetaTrader symbols the volume is tick volume, which is what the volume factor and the volume filter use.

Whichever your broker offers; the signals, scores, levels and dashboard values are the same. MetaTrader 5 draws a filled cloud; MetaTrader 4 draws it with histogram bars. Both builds are in the download.

M15 to H4 is where it is most at home, and every chart on this page was taken between M15 and H4. The timeframe rows default to M15, H1 and H4, so on a daily chart they look at lower timeframes; change them under Multi-Timeframe.

Anything you do not use, without changing a single signal: Show Score on Signals shortens the plates, Show Breakout Labels hides the breakout markers, Show S/R Zones the boxes, Show TP/SL Labels the price tags, and Signal Label Size shrinks the plates. Turning a signal type off in Signal Types removes it completely. Signal Label Cap (450 by default) limits how many plates stay on the chart.

Yes. Each chart runs its own copy with its own dashboard, levels and alerts, and every alert names its symbol and timeframe. Two copies on the same chart, with different settings, keep their drawings apart.

No. It is an indicator, not an Expert Advisor: it scores, draws the cloud and the levels, shows the dashboard and raises the alerts. You place and manage the orders, with whatever position size and risk your account allows.

Checkout takes you straight to a private download page with one zip holding both builds and both manuals. Bookmark that page for re-downloads.

Yes. Free updates to the current version are posted on the same download page.

No. The download holds the compiled .ex4 and .ex5 files and the two manuals.

A 60-day money-back guarantee. If Kumo Forge Pro is not for you within 60 days of purchase, e-mail [email protected] for a full refund, no questions asked.