The combination of trend following and equities is not additive. It is geometrically synergistic: a blended portfolio produces terminal wealth greater than either component alone, with a maximum drawdown smaller than either component alone. This is what happens when two negatively correlated, oppositely skewed return streams compound together.
A 60/40 blend of the S&P 500 and the TF Index turns $100,000 into $820,776. The S&P 500 alone turns it into $747,697. The TF Index alone turns it into $667,058. The blend produces more wealth than either component. At the same time, its maximum drawdown is 24.8%, far below the S&P 500’s 50.9% and only modestly above the TF Index’s 21.0%. More return. Less drawdown. This is not a statistical trick. It is geometric synergy, and the rest of this episode explains why it happens.
The previous episodes built the components. Episode 2 established the multiplicative mathematics of compounding. Episodes 3 and 4 showed that the TF Index produces superior geometric efficiency over 26 years. Episodes 5 through 7 explained the mechanism: why markets trend, how the cut protects the engine, and how letting winners run extends the right tail. Episode 8 demonstrated crisis alpha, the property that makes trend following’s returns most valuable precisely when equities are most destructive.
This episode assembles those components into a portfolio. It does not ask which is better, equities or trend following. It asks what happens when they are combined. The answer is geometric synergy: the whole exceeds the sum of its parts because the components interact in a way that reduces the geometric drag of drawdowns while preserving the geometric fuel of compounding.
The Geometric Efficient Frontier
The traditional efficient frontier, introduced by Harry Markowitz in 1952, plots risk against return using volatility as the measure of risk. This is an arithmetic framework. It treats upside and downside volatility as equally undesirable. It assumes returns are normally distributed. It optimises for a single period rather than for the multi-period compounding that actual investors experience.
The geometric efficient frontier replaces volatility with maximum drawdown and replaces arithmetic return with compound annual growth rate. It optimises for what investors actually care about: how much terminal wealth the portfolio creates and how deeply it falls along the way. The MAR ratio, CAGR divided by maximum drawdown, measures the geometric efficiency of each allocation.
The following table shows blended portfolios of the S&P 500 and the TF Index at every 10% allocation increment, rebalanced monthly, from January 2000 to January 2026.
TABLE 9.1 — S&P/TF allocation grid: terminal value, CAGR, max drawdown, MAR, Sharpe, skewness, and GFC / 2022 returns at each 10% increment
The table reveals three distinct optima depending on the investor’s objective. The maximum terminal value occurs at 60/40 (S&P/TF), producing $820,776 with a CAGR of 8.41%. The maximum MAR ratio occurs at 40/60, producing a MAR of 0.702 with a maximum drawdown of just 11.8%. The maximum Sharpe ratio occurs at 50/50, producing 0.67. All three optima sit between 40% and 60% TF allocation. The data is not ambiguous about the direction: a substantial trend following allocation improves every geometric metric in the table.
Chart 9: The Geometric Efficient Frontier. Each point is a different S&P/TF allocation, coloured by TF percentage. Higher and to the right is better (more CAGR, less drawdown). The 60/40 bond portfolio (diamond) sits well below the equity/TF frontier. Allocations between 40–60% TF dominate all others.
Geometric Synergy: Why the Whole Exceeds Its Parts
The most striking feature of the table is not the location of the optima. It is that every blend between 10% and 80% TF allocation produces a terminal value greater than either the S&P 500 alone or the TF Index alone. In an arithmetic world this cannot happen. If two assets are simply averaged, the blend’s return must lie between the two components. But compounding is multiplicative, not additive, and in a multiplicative world a blend can exceed both of its components.
The source of that excess is drawdown compression, and it is essential to be precise about where the compression comes from, because the answer is the difference between a robust structural claim and a fragile one.
Drawdown compression is a contemporaneous effect. At the moment equities collapse, the trend following sleeve is rising. The two sleeves move in opposite directions at the same instant, so the combined account value falls far less than the equity sleeve alone. No trade is required for this to happen. It is present in a blend that is purchased once and never touched again. It is a property of the negative crisis correlation between the two return streams, not of any decision made during the crisis.
This is the entire geometric mechanism. The S&P 500’s 50.9% maximum drawdown imposes a punishing tax on compounding: it must earn 104% simply to return to its prior peak, and every month spent recovering is a month not spent building new wealth. The blend’s worst drawdown, a 24.8% peak-to-trough decline during the global financial crisis, requires only 33% to recover. That compression of the left tail is not a smoother ride for its own sake. It is recovered time. The wealth not surrendered in the drawdown compounds forward through every subsequent month, and it is this preserved compounding base, rather than any superior arithmetic return, that carries the blend past both of its components. The blend’s arithmetic return is in fact lower than the S&P 500’s. Its terminal value is higher. That inversion is the whole of geometric synergy: reducing the geometric drag of drawdowns produces more terminal wealth than maximising arithmetic return.
Now the subtlety, and it is the objection a careful reader should raise. The blended portfolios in the table are rebalanced monthly, which resets the weights to target at each month end. During a crisis, when the trend sleeve has surged and the equity sleeve has fallen, that rebalance sells trend following and buys equities; when equities later recover, the portfolio benefits from having bought them cheap. It is tempting to credit this trade with the outperformance. It would be a mistake to do so, for two reasons.
First, rebalancing and trend following are wagers on opposite properties of price. Trend following, inside the manager’s book, exploits persistence: today’s direction tends to continue, so the strategy lets winners run. Rebalancing across the two sleeves does the reverse. It trims whatever has risen and adds to whatever has fallen. It is a contrarian, mean-reverting operation, and it is rewarded only when the moves it leans against subsequently reverse. Under the sustained trends that this series argues are the structural feature of markets, systematically trimming the sleeve that is running is not obviously a gain. It can be a cost. The rebalancing bonus that diversification theory promises is, at bottom, a bet on reversion, and it is in direct tension with the heuristic that the rest of this series defends: cut losses short, let profits run.
Second, the data settles the question. Episode 14 will show that less frequent rebalancing produces higher terminal wealth than monthly rebalancing. If rebalancing less often makes more money, then monthly rebalancing, over the full 26-year record, is on net detracting from terminal wealth rather than adding to it. The monthly rebalance is therefore not the source of the synergy. If anything, it is a modest drag on it.
This is the opposite of a weakness in the argument. It means the geometric synergy does not depend on a particular trading schedule, on buying the dip, or on any contrarian timing. It depends only on the negative crisis correlation between the two return streams, a correlation that compresses drawdowns whether the portfolio is rebalanced monthly, annually, or never. The synergy is structural. It survives the removal of the very trade one might naively credit for it.
To see the compression directly, consider the global financial crisis. At 100% equities, the portfolio lost 50.9% and needed 104% to recover. At 60/40 S&P/TF, it lost 23.5% and needed 31%. At 40/60, it lost 5.7%, needing only 6%. At 30/70, it gained 4.4% and required no recovery at all. Every percentage point of drawdown avoided was compounding energy released into every month that followed. The blend does not smooth the ride for comfort. It shortens the recovery, and a shorter recovery is more time spent compounding forward.
The Visual Evidence
The equity curves make the geometric argument visible.
Chart 10: Equity curves for key S&P/TF blends and the traditional 60/40 bond portfolio, January 2000 to January 2026 (log scale). Shaded regions mark equity crisis periods. The 60 S&P / 40 TF blend (green) finishes above both pure S&P and pure TF, with visibly shallower drawdowns during crises.
Two features carry the chart. The first is the behaviour during the global financial crisis. Between late 2007 and early 2009, a 100% equity investor watched $100,000 fall to roughly $49,000. The 60/40 S&P/TF investor watched the same $100,000 fall to roughly $77,000. The difference is not only numerical, it is behavioural. A 24% drawdown is endurable. A 51% drawdown threatens the investor’s capacity to stay invested at all, and the investor who sells at the bottom converts a temporary loss into a permanent one. Holding is the precondition for the recovery that follows, and the shallower drawdown is what makes holding possible.
The second feature is the traditional 60/40 bond portfolio, which finishes near $520,000: roughly $300,000 below the 60 S&P / 40 TF blend, and below even the pure S&P 500, despite its reputation for lower risk. The bond allocation did not fail because bonds performed badly in isolation. It failed because bonds offer none of what trend following offers during a crisis: positive skew, crisis alpha that generates double-digit returns during equity collapses, and a negative equity correlation that intensifies precisely when it matters most. The 40% held in bonds was working at a fraction of the geometric efficiency that the same 40% in trend following would have delivered.
The MAR Frontier
The MAR ratio measures geometric efficiency: how much CAGR the portfolio earns per unit of maximum drawdown endured. A higher MAR means the portfolio is extracting more compounding from less geometric risk.
Chart 11: MAR ratio across all S&P/TF allocations. The ratio rises steadily as TF allocation increases from 0% to 60%, peaking at 0.70 (40 S&P / 60 TF) before declining as the portfolio becomes dominated by TF. The peak MAR is 4.5x the pure equity MAR.
The MAR ratio increases nearly monotonically from 0.157 at 100% equities to a peak of 0.702 at 40/60 S&P/TF. The peak is 4.5 times the pure equity MAR. It then declines as the portfolio becomes dominated by trend following and loses the benefit of the equity component’s long-term growth. The sweet spot, the range where geometric efficiency is maximised, sits between 40% and 60% TF allocation.
This is a striking finding. The geometrically optimal allocation to trend following is not a small diversifying position of 5% or 10%. It is a substantial allocation of 40% to 60%. The benefits of crisis alpha, drawdown compression, and positive skew are proportional to the size of the TF allocation, and they do not peak until the allocation reaches 40 to 60% of the portfolio.
Replacing 60/40
The traditional 60/40 portfolio, 60% equities and 40% bonds, has been the default allocation for balanced investors for decades. Its premise is that bonds provide a counterweight to equity risk: when equities fall, bonds rally, stabilising the portfolio. That premise held, imperfectly, for much of the post-2000 period. It failed in 2022, when stocks and bonds fell together.
TABLE 9.2 — Traditional 60/40 (S&P/bonds) vs. 60/40 (S&P/TF): terminal value, CAGR, max drawdown, MAR, GFC return, 2022 return
Replacing the 40% bond allocation with a 40% TF allocation increases terminal wealth by roughly $300,000, a 58% improvement. It raises the CAGR by 188 basis points. It reduces the maximum drawdown from 32.4% to 24.8%. It lifts the MAR ratio by roughly two-thirds, from about 0.20 to about 0.34. It improves performance during the GFC by 9 percentage points. And it reduces the 2022 loss from 20.1% to 7.8%, cutting that drawdown by more than half.
The 60/40 bond portfolio is not a bad portfolio. It is an incomplete one. It provides some crisis protection, some diversification, and some income. But it provides none of the properties that transform the geometry: positive skew, crisis alpha that turns positive during equity collapses, and the structural negative correlation that intensifies precisely when it is needed. The bond allocation smooths the ride. The trend following allocation compounds.
The Skew Transition
One of the most revealing columns in the allocation table is skewness. At 100% equities, the portfolio’s skew is −0.51, strongly negative. As the TF allocation increases, the skew migrates steadily toward zero and then turns positive. The crossover occurs between 50% and 60% TF allocation. At 40/60 S&P/TF, the skew is +0.01, essentially symmetric. At 30/70, it is +0.07, positively skewed.
This migration is geometrically significant. Episodes 2 and 7 established that positive skew is the return shape compounding rewards: it concentrates variance in the right tail, where the geometric penalty is low, and truncates the left tail, where the geometric penalty is convex. A 100% equity portfolio has close to the worst possible skew for compounding. Each incremental allocation to trend following improves it, and by the time the allocation reaches 40 to 60% TF, the return distribution has been reshaped from geometrically hostile to geometrically efficient.
The skew transition also reveals the nature of the combination. This is not a case of blending two return streams with similar properties. The S&P 500 and the TF Index have opposing distributional characteristics. One is negatively skewed, the other positively skewed. One has its worst months during crises, the other its best months during crises. The blend does not average these properties. It neutralises the equity portfolio’s structural weakness by introducing the trend following process’s structural strength. The negative skew of equities is offset by the positive skew of trend following. The left tail of one is counterbalanced by the right tail of the other. The result compounds more efficiently than either component.
What This Is Not
The optimal allocation of 40 to 60% to trend following is not a prediction. It is a backtest. The 26-year dataset is long enough to capture multiple regimes, including two major equity bear markets, an inflation shock, a pandemic, and extended bull markets. But it is not a guarantee that the next 26 years will produce the same optimal allocation.
What the data does demonstrate is the direction and the mechanism. The direction is clear: increasing TF allocation from zero improves every geometric metric until the allocation becomes substantial. The mechanism is clear: crisis alpha and drawdown compression reduce the geometric drag that erodes compounding. These are structural properties of the process, not artefacts of a particular time period. The cut will continue to truncate the left tail. The trend ride will continue to extend the right tail. The feedback architecture of markets will continue to produce sustained directional movements. The crisis alpha property will persist because crises will continue to produce the strongest trends.
The specific numbers will vary. The specific optimal allocation will shift. But the geometric argument for a substantial allocation to trend following does not depend on any specific number. It depends on the interaction between crisis alpha and the convexity of the recovery curve, an interaction that is mathematical rather than empirical.
There is also the question of implementation. The TF Index is a composite of many managers and is not directly investable. An investor seeking trend following exposure must choose among individual managers, funds, or replication strategies, each with its own fees, capacity constraints, and tracking differences. The geometric argument establishes the theoretical case. The practical implementation requires additional analysis that lies beyond this episode’s scope.
The Running Ledger
Our $100,000 expands to include the blended portfolios. The geometric efficient frontier reveals the architecture of wealth.
The 60 S&P / 40 TF blend leads all portfolios in terminal wealth except Berkshire Hathaway, which benefits from leverage and exceptional stock selection. The 40 S&P / 60 TF blend leads all portfolios in MAR ratio by a wide margin, achieving a maximum drawdown of just 11.8%, the shallowest in the table. The traditional 60/40 bond portfolio finishes last in terminal wealth and second-to-last in MAR ratio. The geometric argument for replacing bond diversification with trend following diversification is comprehensive.
The Bridge
The evidence is strong. But evidence is not complete without confronting its limitations. Every strategy has costs, and trend following’s costs are real: extended drawdowns, periods of underperformance, whipsaws, the psychological burden of a strategy that loses more often than it wins. The geometric case does not depend on these costs being small. It depends on them being worth paying.
Episode 10 begins Act III: the evidentiary core. It will deploy the NilssonHedge data in full force to show the consistency of the trend following record across 41 independent managers over two decades. It will address the survivorship bias objection, the period-dependence objection, and the question of whether the geometric advantage persists across different implementations. The series shifts from argument to evidence.
Data and Sources
All performance data from the NilssonHedge Trend Following Performance Database (January 2000 to January 2026). All returns are net of management and performance fees. 313 monthly observations. Blended portfolios are constructed using monthly rebalancing: each month’s return is the weighted average of the S&P 500 TR and TTU TF Index monthly returns at the stated allocation. Terminal values are the cumulative product of (1 + monthly blend return) applied to $100,000. Maximum drawdown is the largest peak-to-trough decline in the cumulative return series. Crisis-window returns reported in the text measure the loss over the defined crisis dates and may differ slightly from the full peak-to-trough maximum drawdown. MAR ratio = CAGR / |MaxDD|. Sharpe ratio uses annualised excess return over a 2% risk-free rate divided by annualised standard deviation of monthly returns. The 60/40 bond portfolio uses 60% S&P 500 TR and 40% Vanguard Total Bond Market Index Fund (VBMFX). The geometric efficient frontier concept extends Markowitz (1952) by substituting maximum drawdown for volatility and CAGR for arithmetic mean return. The sensitivity of these results to rebalancing frequency is examined directly in Episode 14. The 2022 return for the 60 S&P / 40 TF blend is estimated from the monthly blend returns during January to September 2022.
Want the theoretical foundation for why trend following works?
The Fractals of Finance: Determinism, Adaptation and the Geometry of Markets
The book explores the full architecture of feedback, fat tails, and fractal structure in financial markets, and what it means for how we trade, invest, and understand risk.
Available now on Amazon in paperback, hardcover, and Kindle.
Want a practical field manual for trading trends and capturing outliers?
The Aussie Turtles Trend Following Guide: A Field Manual for Hunting Outliers adapts the timeless principles of the original Turtle traders into a systematic, rules-based approach for modern markets. Co-authored with Adam Havryliv.
Available now on Amazon in paperback, hardcover, and Kindle.