Ichimoku Cloud
A multi-component system that defines trend, momentum, support, and resistance in a single visual framework using five lines.
Overview
Ichimoku Kinko Hyo ("one glance equilibrium chart") was developed by Japanese journalist Goichi Hosoda over 30 years of research and published in 1969. It combines five lines to give a complete picture of trend direction, momentum, and key support/resistance levels — all in one chart. The shaded "cloud" between two of its lines is its most distinctive feature, showing the balance of power between buyers and sellers.
How it looks on a chart
Illustration only — synthetic data generated for visual reference.
Ichimoku may look complicated at first, but each part tells you something specific. The cloud (Kumo) is the most important element: when price is above the cloud, the market is in an uptrend. When price is below the cloud, it is in a downtrend. When price is inside the cloud, the market is in a transition or consolidation phase. The two cloud lines — Senkou Span A (green) and Senkou Span B (red) — form the cloud boundary. A thick cloud means strong support or resistance. A thin cloud means weaker support/resistance that is more easily broken. The cloud is plotted 26 periods ahead, so it serves as a visual forecast of future support and resistance. The Tenkan-sen (Conversion Line) and Kijun-sen (Base Line) work like moving averages. When Tenkan crosses above Kijun, it is a bullish signal. The Chikou Span (Lagging Span) is today's closing price plotted 26 bars in the past — if it sits above the price from 26 bars ago, the current momentum is considered positive.
The five components use midpoint calculations rather than averages: Tenkan-sen = (9-period High + 9-period Low) / 2. This midpoint approach reflects equilibrium price — the middle of the range — rather than mean price. Kijun-sen uses 26 periods; Senkou Span A = (Tenkan + Kijun) / 2 plotted 26 periods forward; Senkou Span B = (52-period High + 52-period Low) / 2 plotted 26 periods forward; Chikou = Close plotted 26 periods back. The original parameters (9, 26, 52) were developed for Japanese stock markets that traded 6 days a week (26 sessions = 1 month). For modern 5-day markets, some analysts adjust to (7, 22, 44) or (10, 30, 60) for crypto. For intraday charts, parameter scaling is debated — many practitioners retain the original values on hourly charts. High-quality signals require all components to align: price above cloud, cloud is green (Span A > Span B), Tenkan above Kijun, Chikou above historical price and above cloud. This "perfect bullish alignment" is rare but provides high-conviction entries. Partial alignment is more common and requires additional context.
Ichimoku is a complete trading system designed for discretionary analysis, which makes systematic implementation non-trivial. The cloud twist (where Span A crosses Span B) is plotted 26 bars in the future — a look-ahead feature in its original form. In backtesting, the future cloud must be converted to a signal at the point it was "visible," which requires careful time-shifting logic. Research on Ichimoku in systematic settings is limited compared to simpler indicators. Murphy (2009) and more recent studies on Asian equity markets find that Ichimoku-based strategies achieve positive risk-adjusted returns on trending assets but suffer significantly in ranging markets, similar to other trend systems. The five-component nature makes it sensitive to parameterization. Gilito's engine decomposes Ichimoku into its individual signals and tests each sub-signal (Tenkan/Kijun cross, cloud breakout, Chikou confirmation) separately and in combination. This decomposition reveals that the cloud breakout signal alone carries most of the predictive value, while the Chikou confirmation adds marginal improvement at the cost of additional lag.
Formula
Tenkan-sen = (High₉ + Low₉) / 2 Kijun-sen = (High₂₆ + Low₂₆) / 2 Senkou A = (Tenkan + Kijun) / 2, plotted +26 Senkou B = (High₅₂ + Low₅₂) / 2, plotted +26 Chikou = Close, plotted −26
- 1.Calculate Tenkan-sen: (highest high + lowest low) over last 9 periods, divided by 2.
- 2.Calculate Kijun-sen: (highest high + lowest low) over last 26 periods, divided by 2.
- 3.Compute Senkou Span A = (Tenkan + Kijun) / 2; plot it 26 periods into the future.
- 4.Compute Senkou Span B = (52-period high + 52-period low) / 2; plot it 26 periods into the future.
- 5.Plot Chikou Span = today's close, shifted 26 periods back on the chart.
Parameters
| Parameter | Default | Range | Description |
|---|---|---|---|
| Tenkan Period | 9 | 5–20 | Conversion line lookback period. |
| Kijun Period | 26 | 15–52 | Base line lookback period. |
| Senkou B Period | 52 | 30–104 | Slow cloud boundary lookback period (typically 2× Kijun). |
Trading signals
bullish: Price breaks above the cloud with Span A > Span B
Strong bullish breakout — cloud acting as support, trend confirmed up.
bearish: Price breaks below the cloud with Span B > Span A
Strong bearish breakdown — cloud acting as resistance, trend confirmed down.
bullish: Tenkan-sen crosses above Kijun-sen above the cloud
Strong TK cross (Golden Cross in Ichimoku) — high-quality long signal.
bearish: Tenkan-sen crosses below Kijun-sen below the cloud
Strong TK cross (Dead Cross in Ichimoku) — high-quality short signal.
Limitations
- •Complex system with five components — difficult to systematize without decomposing signals.
- •Original parameters designed for 6-day Japanese market weeks; direct application to Western or crypto markets may not be optimal.
- •Future cloud plotted 26 periods ahead creates look-ahead bias issues in naive backtesting implementations.
- •All five components aligning simultaneously is rare, leaving many periods without a clear signal.
Gilito decomposes Ichimoku into its five individual signal components and backtests each independently and in combination, allowing it to identify which elements (cloud breakout, TK cross, Chikou confirmation) contribute the most alpha on each specific asset class. Parameter sweeps over Tenkan, Kijun, and Senkou B periods are run against out-of-sample data to find robust configurations.