Ichimoku Kinko Hyo

Definition

Ichimoku Kinko Hyo (“one-glance equilibrium chart”, developed by Goichi Hosoda, published 1969) is a five-line system built from rolling range midpoints — not averages of closes — with two lines projected FORWARD to form the “cloud” (kumo) and one plotted BACKWARD. It bundles trend state, momentum, and support/ resistance conventions into one display. Structurally it is Donchian-midline machinery with time displacement.

How it works / structure

• Components: Tenkan-sen = 9-bar range midpoint ((H+L)/2); Kijun-sen = 26-bar midpoint; Senkou A = average of those two, plotted 26 bars ahead; Senkou B = 52-bar midpoint, plotted 26 ahead (A and B bound the cloud); Chikou = today’s close plotted 26 bars back.
• Signal conventions (engine-executable, each pinned separately): price above/below cloud (trend state); Tenkan/Kijun cross (fast signal); cloud color/twist; Chikou clear of past price. The traditional 9/26/52 windows derive from the Japanese 6-day trading week — calendar archaeology, not optimization.
• What it actually is: range midpoints ≈ Donchian midlines (indicator-donchian-channels), so Ichimoku states correlate strongly with MA/channel trend states; the forward displacement is presentational — no future information exists, the lines are today’s values drawn at tomorrow’s coordinate.

When it applies

Trend-state definition for practitioners fluent in its vocabulary; multi-condition trend filters (all-conditions- aligned as a conservative trend gate). The platform maps Ichimoku states onto trend_state and treats the five-line bundle as one more trend-family parameterization — not an independent information source.

Risk profile & failure modes

• Complexity theater: five lines and a cloud imply more information than the two range midpoints they contain; conviction scales with visual density, information does not.
• Convention explosion: dozens of documented signal combinations — a wide snooping surface (lens-quantitative).
• Displacement confusion: forward-plotted lines are routinely misread as forecasts; the cloud ahead is history shifted right.
• Range-regime whipsaw: inherited from every trend-family tool.

Evidence & limits

Peer-reviewed evidence on Ichimoku specifically is scant; a handful of market-specific studies show mixed results, and it falls under the technical-rule survey verdicts (Park-Irwin 2007: mixed, snooping-fragile). Since its components reduce to range midpoints, the honest evidence statement is: Ichimoku inherits the trend-following evidence base (strategy-momentum, strategy-futures-trend-following) with zero documented incremental edge for the extra machinery. Cloud-specific lore is folklore, labeled.

Falsifiable-thesis examples

Illustrations only, not signals:

• “X, closing above its cloud with Tenkan > Kijun, will outperform its sector over the next month” — falsified by the pair’s returns.
• “Ichimoku all-aligned trend state adds replay Sharpe over a 50/200 SMA state on universe U this year” — falsified by the paired replay.

Cross-references

• Structural relatives: indicator-donchian-channels (range midpoints), indicator-sma (trend states)
• The evidence it inherits: strategy-momentum, strategy-futures-trend-following
• Platform binding: trend_state
• Method caveats: lens-technical, lens-quantitative