The 84-Week Cycle: Why Hurst's 18-Month Nominal Keeps Showing Up

When you run cycle analysis on enough markets, patterns start to repeat. Not vague similarities. The same specific cycle lengths, appearing with statistical significance across completely unrelated instruments. The 84-week cycle is the most consistent example we have found.

We ran full spectral analysis on the US Dollar Index (weekly, all available bars back to 2008) and found an 84-week cycle at 100% relative strength, the strongest cycle in the entire spectrum. That alone would be interesting. What makes it significant is that this same cycle keeps appearing, in equities, commodities, crypto, and currencies, often as one of the top three strongest detected periods.

If you are familiar with J.M. Hurst's nominal model, this number should look familiar. It is his 18-month cycle.

What the Data Shows

Using FractalCycles' Goertzel-based spectral analysis on weekly data with all available bars, the 84-week cycle (give or take a few weeks, consistent with Hurst's Principle of Variation) registers as one of the dominant periods across multiple asset classes:

• US Dollar Index (DX-Y.NYB): 84-week cycle at 100% relative strength, the single strongest detected cycle in the entire spectrum. Currently in a Rising phase.
• S&P 500: The 80-90 week range consistently appears as a high-strength cycle, aligning with the intermediate moves that define bull and bear market legs.
• Bitcoin: Despite being a newer asset class, BTC exhibits a strong cycle in the 80-84 week range that corresponds with its intermediate trend reversals.
• Gold: The precious metals complex shows this cycle prominently, consistent with gold's role as a macro-driven asset responsive to institutional rebalancing.
• Crude Oil: Energy markets, driven by supply cycles and institutional positioning, register strong spectral peaks in this same range.

This is not a coincidence. When the same period appears across uncorrelated assets, it points to something structural in how markets operate.

Mapping to Hurst's Nominal Model

In the 1970s, J.M. Hurst identified a hierarchy of cycle periods that appear across financial markets. His nominal model includes:

• 18-year cycle: The secular cycle
• 54-month cycle: Roughly aligned with business cycles
• 18-month cycle (~78-84 weeks): The intermediate cycle
• 40-week cycle: Key swing trading cycle
• 20-week cycle: The most commonly traded cycle
• 80-day cycle: Short-term trading cycle
• 40-day, 20-day, 10-day: Progressively shorter cycles

The 84-week cycle we detect maps directly to Hurst's 18-month nominal. The match is not approximate. 18 months at roughly 4.33 weeks per month equals 78 weeks. With Hurst's Principle of Variation (real cycles fluctuate around their nominal period), an observed 84-week cycle is exactly what you would expect from an 18-month nominal.

But the correspondence goes deeper. When we look at the full spectrum from the Dollar Index analysis, other detected cycles also align with Hurst's hierarchy:

• 368 weeks detected (~7 years): Close to the 54-month nominal doubled, suggesting a higher-order cycle
• 183 weeks detected (~3.5 years): Near the 54-month nominal cycle (4.5 years), within the expected variation range
• 102 weeks detected (~2 years): An intermediate harmonic between the 18-month and 54-month cycles
• 70 weeks detected: Close to twice the 40-week nominal
• 40-43 weeks detected: Direct match to Hurst's 40-week cycle

The Goertzel algorithm does not know about Hurst's model. It scans the data mathematically and reports what it finds. The fact that it independently recovers the same hierarchy Hurst identified through manual analysis over fifty years ago is a strong validation of both the method and the model.

Why the 18-Month Cycle Matters

Among Hurst's nominal cycles, the 18-month period holds a special position. It sits at the intersection of several important dynamics:

• Institutional rebalancing: Large funds typically review and rebalance allocations on quarterly and semi-annual cycles. The 18-month period captures the larger rhythm of these institutional flows.
• Central bank policy cycles: Monetary policy operates in cycles that often align with the 12-24 month range. Rate hike cycles, pause periods, and easing cycles create oscillations that the 18-month cycle captures.
• Economic data cycles: Leading economic indicators, inventory cycles, and credit cycles have periodicities that cluster in the 12-24 month range.
• Behavioral time horizons: The 18-month period sits at the boundary between what most participants consider "medium-term" and "long-term," creating a natural inflection point where positioning shifts.

For position traders and investors, the 18-month cycle is arguably the most actionable cycle in Hurst's model. It is long enough to capture significant price moves (Hurst's Principle of Proportionality states that amplitude scales with period) but short enough to provide multiple trading opportunities within a typical investment horizon.

Hurst's Principles in Action

The 84-week finding illustrates several of Hurst's core principles working together:

Principle of Commonality

The same cycle period appearing across currencies (Dollar Index), equities (S&P 500), crypto (Bitcoin), and commodities (Gold, Crude) confirms Hurst's observation that markets share common periodicities. The specific amplitudes differ (the Dollar Index oscillates differently than Bitcoin), but the timing structure is shared.

Principle of Synchronicity

In the Dollar Index analysis, the 84-week cycle currently in a Rising phase coincides with the 40-week and 70-week cycles also in Rising phase. This multi-timeframe alignment creates stronger directional moves. When multiple Hurst nominal cycles point the same direction, the resulting trend is more persistent. When they disagree, you get the choppy, range-bound conditions that frustrate directional traders.

Principle of Nominality

The detected cycles relate to each other by factors close to 2: 40 weeks to 84 weeks is roughly 1:2. 84 weeks to 183 weeks is roughly 1:2.2. This harmonic nesting is exactly what Hurst described. Each cycle nests within the one above it, creating the layered structure that defines market behavior at every scale.

The Variation Problem: Why Most Mathematical Models Fall Short

There is a legitimate criticism leveled at mathematical cycle detection tools, and it deserves a direct answer. Many spectral analysis platforms treat cycles as fixed, rigid periodicities. They detect a cycle at, say, 84 weeks, and then project it forward as if it will repeat at exactly 84 weeks forever. This approach violates Hurst's Principle of Variation, which states that real cycles breathe. A nominal 18-month cycle might measure 76 weeks in one instance, 88 weeks in the next, and 81 weeks after that. The period is alive, not mechanical.

This is a fair criticism of static cycle models. If a tool locks onto a fixed wavelength and projects it forward like a metronome, it will eventually drift out of phase with the actual market. Markets are not clocks. Cycles compress and expand in response to changing conditions, shifting volatility regimes, and the interaction of cycles at different scales.

FractalCycles was built with this understanding. Our approach is dynamic, not static:

• Reanalysis, not extrapolation. Rather than detecting a cycle once and projecting it indefinitely, users rerun analysis as new data arrives. Each analysis reads the current state of the market fresh, capturing the cycle as it actually is, not where a fixed model says it should be.
• Phase tracking over period locking. The Cycle Spectrum table shows the current phase (Rising, Peaking, Falling, Bottoming) of each detected cycle. This phase reading comes from the data itself, not from counting forward from a historical trough. If the 84-week cycle compresses to 78 weeks or stretches to 90 weeks, the phase tracking adapts because it is reading the actual oscillation, not a rigid projection.
• Multiple timeframe awareness. By running analysis at different bar counts and timeframes, you see how cycle lengths shift over time. A cycle that measured 84 weeks over the full dataset might measure 78 weeks in the most recent 400 bars. That variation is information, not error. It tells you the cycle is compressing, which often precedes a stronger reversal.
• Regime context via the Hurst exponent. The Hurst exponent provides a meta-layer that static models lack entirely. When H shifts from 0.55 to 0.48, it signals that the market regime is changing and cycle behavior will change with it. Static models have no mechanism to account for this. FractalCycles surfaces it automatically.

We deliberately refused to join the static mathematical cycle crowd. Cycle analysis that treats markets like clockwork is destined to fail because markets are not clockwork. They are complex adaptive systems where cycles compress, expand, and interact in ways that only a dynamic, re-analysis-driven approach can track. Hurst understood this in the 1970s. His Principle of Variation was not an afterthought. It was central to his framework. Any tool that claims to apply Hurst's model while ignoring his most practical principle is missing the point.

How FractalCycles Detects Hurst Cycles

Traditional Hurst cycle analysis involved manually fitting envelopes to price charts, a process that was subjective and time-consuming. Modern spectral methods automate this entirely.

FractalCycles uses a pipeline that maps directly onto the Hurst framework:

1. Spectral analysis via Goertzel algorithm: Scans the price data across all possible periodicities and identifies which cycles are present. This is the automated equivalent of Hurst's manual frequency identification.
2. Bartels significance testing: Each detected cycle is tested for statistical significance. This prevents overfitting by filtering out cycles that could be random noise. Only cycles that pass the significance threshold are reported.
3. Hurst exponent calculation: The platform computes the Hurst exponent using multiple methods (R/S analysis and DFA) to determine whether the current market regime is trending, mean-reverting, or random walk. This regime context tells you how to interpret the detected cycles.
4. Composite wave projection: Selected cycles are combined into a forward-looking composite waveform. When you select the 84-week cycle along with the 40-week and other significant cycles, the composite projects where synchronized turning points are likely to occur.

The result is that you can perform a complete Hurst-style cycle analysis on any market in seconds, with statistical validation that manual methods cannot provide.

What This Means for Hurst Cycle Traders

If you already practice Hurst cycle analysis, these findings should be reassuring. The nominal model is not an artifact of 1970s data. The same cycles Hurst identified are still active today, detectable with modern algorithms, and verifiable with statistical tests.

Specifically, here is how to use the 84-week cycle in practice:

• Identify the phase: Is the 18-month cycle currently Rising, Peaking, Falling, or Bottoming? This tells you the intermediate trend direction. In the Dollar Index as of late March 2026, the 84-week cycle is Rising.
• Check for synchronicity: Are shorter cycles (40-week, 20-week) aligned with the 18-month cycle? Aligned cycles produce stronger moves. Opposing cycles produce chop.
• Use the composite projection: Select the 84-week cycle along with other significant cycles and examine where the composite wave projects the next turning point. This is the modern equivalent of Hurst's projected trough dates.
• Validate with the Hurst exponent: A Hurst exponent above 0.55 suggests the market is trending (cycles are producing directional moves). Below 0.45 suggests mean reversion (cycles are producing range-bound oscillation). Near 0.50 suggests random walk behavior where cycles have less predictive value.

Beyond the 18-Month Cycle

The 84-week finding is one data point in a broader picture. The entire Hurst nominal hierarchy, from the 10-day cycle up through the 54-month and 18-year cycles, can be detected using the same spectral methods at different timeframes.

Running analysis on daily data reveals the shorter Hurst nominals (10-day, 20-day, 40-day, 80-day). Weekly data, as we used for the Dollar Index, captures the intermediate cycles (20-week, 40-week, 18-month). Monthly data extends into the longer periodicities (54-month and beyond).

The key insight from Hurst, confirmed by modern spectral analysis, is that these cycles are not independent. They nest. The 20-day cycle oscillates within the 40-day cycle, which oscillates within the 80-day cycle, and so on up the hierarchy. Understanding this nesting structure is what transforms isolated cycle readings into a complete market picture.

Running Your Own Analysis

To verify these findings on any market:

1. Navigate to the Live Chart in your FractalCycles dashboard
2. Search for any symbol (DX-Y.NYB for Dollar Index, SPY for S&P 500, BTC-USD for Bitcoin)
3. Set the timeframe to Weekly
4. Click Run Analysis
5. On the analysis page, change the Bars dropdown to All to use the full dataset
6. Look at the Cycle Spectrum table. Sort by Strength to find the dominant cycles. Look for periods in the 78-90 week range. That is your 18-month Hurst nominal.
7. Check the shorter cycles too: 38-42 weeks maps to Hurst's 40-week nominal, 18-22 weeks maps to the 20-week nominal

The spectral analysis will objectively show you which of Hurst's nominal cycles are currently active in your chosen market, their relative strengths, and their current phase. No manual envelope fitting required.

Conclusion

J.M. Hurst identified the 18-month cycle over fifty years ago using manual calculations on graph paper. Today, the Goertzel algorithm independently detects the same 84-week period across currencies, equities, commodities, and crypto, confirming both the persistence of the cycle and the validity of Hurst's nominal model.

For Hurst cycle practitioners, FractalCycles provides the modern toolset to apply the framework you already understand. For traders new to cycles, this is the entry point: start with the 18-month cycle, learn to read its phase, and build outward from there.

The cycles Hurst discovered have not changed. The tools to detect them have.