Dual Momentum Pivot5: the deep dive

A rules-based tactical strategy over nine asset-class ETFs. Every month it ranks them by trend, holds the five strongest at equal weight, and moves any whose trend has turned negative into cash. Over 1991–2026 the backtest compounded at 10.6% a year versus 7.8% for global stocks, at roughly half the volatility and less than a third of the worst drawdown — bought by giving up upside in sharp equity rallies, higher turnover, and the tax drag of frequent trading. This is the full guide: why a tactical layer, how it’s built, the evidence, why it works, when it fails, and who it’s for.

Part I — Why a tactical layer

“Buy the S&P and chill” is riskier than you think

For most of the 2000s, one belief sat under almost every money decision Americans made: house prices don’t fall. It hadn’t happened across the country in living memory, so it stopped feeling like a belief and started feeling like a fact. Banks lent as if prices could only rise; Wall Street built products on the same assumption. Then it broke, and because everything was built on it, everything fell at once.

Investing has a linchpin of its own right now, and it’s “buy the S&P and chill.” It’s not that investors think stocks can’t fall — it’s that almost nobody questions whether a concentrated stock portfolio is the right starting point. For decades it’s been right: US stocks won everything. So, like house prices in 2006, it stopped looking like a bet and started looking like a fact.

But if risk is the probability of not meeting your long-term investment goals (my favourite definition), then “buy the S&P and chill” quietly demands three things most investors don’t have:

If you genuinely have all three, you can probably stop reading — buy-and-hold is a great strategy for you. Otherwise, as Yogi Berra put it: “In theory, there’s no difference between theory and practice. In practice, there is.” And today widens that gap: fifteen exceptional years for US equities have bred unrealistic expectations, valuations leave little margin for error, and enough is changing fast that something is likely to go wrong.

There’s a better way

Not better in the absolute — better for you. For fifteen years, first running risk parity and alternative risk premia at large London funds and then investing my own money, I’ve been obsessed with one question: how do I maximise my chances of reaching my target returns, even if the future unfolds very differently from what I expect? The process I landed on has three layers:

The InvestLab framework — strategic core plus tactical layer plus satellites. The structure is what matters; this guide is one building block from the middle layer. Source: InvestLab (illustrative).

This guide is one of those building blocks: a single tactical strategy from the middle layer. Rules-based, evidence-backed, fifteen minutes a month. It’s intentionally detailed — you don’t need to read it cover to cover, but if you want to understand a strategy properly you have to go deeper. Start with the idea and the rules, skim the results, and jump to the conclusion.

Part II — The strategy

Meet Dual Momentum Pivot5

The intuition is simple: picking from a robust universe of assets, you hold what’s working in the current environment and step away from what isn’t. As conditions change, the portfolio adapts — leaning into the assets being rewarded and scaling back the rest. Dual momentum draws on the oldest, most robust anomaly in markets: price trends persist. Assets that have been rising tend to keep rising; bear markets are often long and protracted. Not forever, not always, but often enough — across 137 years of data covering 67 markets (Hurst, Ooi & Pedersen, AQR, 2017) — to be the basis of a system.

The strategy uses two ideas in sequence. It starts with relative momentum: ranking the asset classes against each other to see which are strongest right now. But sometimes almost everything is falling at once, and the “least bad” option still isn’t good enough. That’s where absolute momentum comes in — it asks whether an asset is actually trending up. If not, the strategy steps aside into cash and waits. It doesn’t predict the wind; it adjusts the sails to the wind that’s already blowing.

The universe: nine assets across regimes

A momentum strategy is only as good as the universe it runs on. We want assets we’d be happy holding in a buy-and-hold portfolio, that cover the major macro regimes (so something is always trending), with deep, liquid, low-cost histories — kept lean to avoid overfitting.

Three equity engines for growth, three bond engines for slowdowns, three real assets for inflation and stress — so something is trending in every regime. We don’t predict the next regime; we make sure we hold assets that can perform across a wide range of them, then let momentum decide which deserve capital:

The four rules

Step 1 — score each asset by its trend. The simplest signal is an asset’s return over the past 12 months. We take a slightly more robust approach and average returns across five lookbacks — 12, 9, 6, 3 and 1 month. Diversification works for signals too: blending lookbacks is less sensitive to any single parameter and captures both shorter- and longer-term trends.

Step 2 — hold the five strongest (relative momentum). Rank all nine, take the top five — the assets the market is currently rewarding. Five is a deliberate balance between concentration and diversification (20% max per asset), not a number tuned to flatter the backtest.

Step 3 — drop the ones actually falling (absolute momentum). This is the “dual” in dual momentum. For each of the top five, check whether its score is positive. If it is, hold it; if not, that 20% slot goes to cash. When markets turn broadly hostile, several assets fail at once and the portfolio becomes genuinely defensive.

Step 4 — size equally, rebalance monthly. Every asset that passes both filters gets 20%; each failed slot’s 20% goes to cash. Equal weights keep it simple, transparent, and hard to overfit. Then rebalance and wait — about fifteen minutes.

Here’s the whole process on one illustrative month — rank by momentum (relative), then keep only those with a positive trend (absolute):

How to read it: relative momentum selects the top five; absolute momentum then checks each has a positive trend — IEF ranks top-five but its trend is negative, so its slot moves to cash rather than buying a falling asset. Illustrative only; scores and tickers are examples, not a recommendation.

Part III — Track record

35 years of evidence

Here’s how a $10,000 investment at the end of 1991 would have grown, for the strategy and the benchmark:

Growth of $10,000 — Dual Momentum Pivot5 vs VT (log scale). The strategy compounded $10,000 to $335,963 versus $129,871 for the global-equity benchmark — roughly 2.6× the terminal wealth, with materially smaller drawdowns along the way. Net of 10 bps round-trip costs.

At first glance this looks like a story about higher returns. It isn’t. The more interesting finding is that those returns came alongside a very different risk profile: roughly half the volatility, less than a third of the maximum drawdown, and more than double the Sharpe. It made money more frequently, the distribution is less negatively skewed, and the tails are thinner — losses tended to be smaller and less frequent, and long-term results more predictable. That combination is rare. Four characteristics stand out.

1. Higher returns

The strategy compounded at 10.6%/yr (after costs) versus 7.8% for VT. That 2.8-point difference may not sound extraordinary, but over 35 years it translated into more than 2.5× the terminal wealth. What’s more interesting is when the outperformance was earned — not steadily, but concentrated in a handful of difficult environments: the dot-com crash, the GFC, and to a lesser extent the 2022 inflation shock.

Relative performance: strategy ÷ VT (1.0 = same growth). Almost all the outperformance was earned when global equities suffered — the dot-com bust, the GFC, and 2022 — when the strategy rotated away from struggling assets into bonds, gold, or commodities.

Five-year stretches: strategy vs VT. It made its money during regime changes (1990–94, 2000–04, 2005–09, 2025–26) and roughly kept pace or lagged when one asset class — usually US equities — dominated uninterrupted (1995–99, 2010–14, 2015–19, 2020–24).

You might expect a strategy like this to lag badly during equity bull markets. Yet it broadly kept pace through the 1990s run-up and the post-GFC recovery, because other assets were pulling their weight — bonds in the 1990s, real estate in the early 2010s. The real challenge was 2015–2024, when US equities were often the only major asset class generating exceptional returns. The most encouraging finding, though, is that over the full backtest returns came from multiple engines rather than a single bet:

Annualised contribution to return by asset (1991–2026, ranked). Gold contributed roughly as much to long-run results as US equities despite a much smaller average allocation — a sign of how well the trend filter timed it. EM, REITs, bonds, and developed-market equities all pulled their weight. Monthly time-integrated arithmetic contribution per asset.

2. Smaller and shorter drawdowns

Returns attract attention, but drawdowns determine whether you’ll stay invested long enough to earn them. The strategy’s maximum drawdown was −16.9%, against VT’s −55.1%. A 17% drawdown turns $100,000 into roughly $83,000 — painful but recoverable. A 55% drawdown leaves it worth under $45,000 — the kind of loss that makes investors abandon the plan.

Drawdowns — peak to trough (%). The strategy’s worst drawdown was −16.9%, less than a third of VT’s −55.1%, and it recovered within roughly 18 months while VT took five years.

It didn’t just lose less — it recovered faster:

Dual Momentum Pivot5 — top 5 drawdowns.

VT — top 5 drawdowns over the same window.

The strategy’s deepest drawdown recovered in roughly eighteen months; the benchmark’s largest required five to seven years. Duration matters almost as much as depth: the longer a portfolio stays underwater, the greater the behavioural pressure to abandon the plan.

3. More predictable outcomes

The strategy’s most underrated feature isn’t its higher return — it’s how much narrower the range of outcomes was at almost every horizon. Nobody experiences the average; you experience one realised outcome that depends largely on when you happened to start.

Rolling 10-year annualised return. The strategy never produced a negative 10-year outcome — even its weakest decade delivered roughly 5% annualised. The benchmark fell as low as −3.4%/yr during the lost decade of 2000–2009.

Rolling 3-year annualised return. Across 384 overlapping windows the strategy never fell below 1.4% annualised. The benchmark was negative in 67 windows (~17%), bottoming at −19.4%/yr in the dot-com bust. The spread was almost half as wide (std 4.9pp vs 8.4pp).

Calendar-year returns (%). Over 35 years the strategy’s worst year was −6.2%; the benchmark’s was −41%. The strategy had only two years worse than −5%; the benchmark had eight. It gave up some spectacular upside years but avoided the disastrous ones.

Distribution of annual returns. Compared to the benchmark, the strategy’s returns were more tightly clustered, less negatively skewed, and had thinner tails — sacrificing some of the right tail for a big improvement in the left.

4. An asymmetric profile

By now a pattern should be emerging. The strategy also exhibits a very different payoff profile from a traditional equity portfolio:

Annual returns: strategy vs benchmark (each dot = one year). The “smile” is the visual signature of the trade-off — when VT was deeply down, the strategy was up; when VT had big up years, it participated but captured less. Smaller upside, much smaller downside.

This “call option” profile comes from three mechanisms working together. First, relative momentum keeps the portfolio invested in whatever is leading — often equities and real assets in bull markets. Second, relative momentum also plays defence, rotating toward bonds, gold, or commodities as leadership changes (in 2022, commodities offset equity losses). Third, when trends deteriorate across the board, the absolute-momentum filter moves part of the portfolio into cash — the mechanism that cuts off the left tail. You can see all three in the holdings:

Average annual weight by asset. The composition shifted dramatically across regimes — heavy in commodities and bonds into the GFC, equity-tilted in the 2010s, diversified across gold, EM and commodities since 2022. The system never sat still for long.

Cash allocation over time (% — defensive mode). Cash is the safety valve, clustering in defensive periods (1992 warm-up, 2008–09, 2015–16, 2022–23). The same mechanism that protected capital also created lag in sharp recoveries like 2009.

An important note on the backtest

Part IV — Why it works

Why it works — and why it should keep working

The strategy is good at both offence (relative momentum) and defence (relative and absolute momentum). But the deeper question is why those signals should work at all. In an efficient market, an effect this well-documented — visible across 215 years and across equities, bonds, currencies and commodities (Geczy & Samonov, 2016) — should have been arbitraged away. It hasn’t. Six reasons keep showing up:

• Economies move slowly. Growth, inflation, rates and earnings evolve over months and years. When an environment starts rewarding one group of assets, that regime often persists long enough for a rules-based process to adapt.
• Investors underreact. People anchor on old beliefs and update too slowly, so prices drift in the direction of news before fully reflecting it.
• Trends become self-reinforcing. Flows, narratives, performance-chasing and herding push a working asset further — often longer than intuition suggests.
• Capital moves slowly. Large investors face mandates, committees, benchmarks and career risk; they can’t instantly rotate, which helps trends persist.
• It’s divergent, not convergent. Value and carry close the gap they exploit as capital piles in; momentum pushes prices further along the trend, so it’s far harder to arbitrage away — at the cost of sharper reversals when trends break.
• And it’s uncomfortable to hold. Momentum gives back a lot, fast, at turning points, and lags exactly when buy-and-hold feels easiest. That discomfort is partly the point — some of the return is compensation for bearing it, which makes it more likely to last.

That doesn’t mean it will always work, or work as well as in this backtest. But the reasons are rooted in economic cycles, human behaviour, institutional constraints, and the risk of being wrong at turning points — a much stronger foundation than a nice-looking backtest.

Part V — Where it breaks

When does the strategy fail?

This may be the most important section. The biggest risk usually isn’t that the strategy fails — it’s that you abandon it when it becomes uncomfortable. Four environments tend to hurt performance:

• Choppy, sideways markets. Momentum needs trends. When prices move back and forth, signals whipsaw. 2015 is the clearest case: −6.2% while global equities were broadly flat. The culprit wasn’t a crisis — it was noise.
• Sharp turning points. Momentum is a lagging signal; it doesn’t catch bottoms. In 2009 it returned 11.8% vs 32.7% for VT; in 2023, 2.7% vs 22.0%. The same defensive mechanisms that protected capital became a headwind in the recovery.
• Fast crashes. It rebalances monthly. Most bear markets unfold slowly enough, but sudden crashes (October 2008, March 2020) can catch it between rebalances.
• Long periods of market concentration. The hardest failure mode psychologically. From 2013–2019 US equities dominated; the strategy compounded at 5.0%/yr vs 9.9% for VT. Nothing was broken — diversification simply became a disadvantage.

Costs, turnover and taxes

There’s also a more mundane drag: turnover. The strategy turns over roughly 211% a year — about two of its five holdings every month. Transaction costs are already in the backtest (reducing returns by ~0.4%/yr), but real-world costs may be higher. Taxes can matter more: rebalancing creates more taxable events than buy-and-hold, so where you can, run it in a tax-advantaged account — tax can knock another 1–2%/yr off returns. That said, it’s usually less punishing than the turnover suggests: the big winners keep trending and stay in the portfolio, compounding tax-deferred, while what’s sold is mostly recent positions with little gain.

Part VI — Is it right for you?

The trade-off and the blend

You give up some upside in strong bull markets, accept higher turnover, and occasionally endure years of frustrating underperformance. In return you should get a smoother ride, materially smaller drawdowns, and a higher probability of reaching your destination without abandoning the plan. Whether that’s a good deal depends on your objective. If your goal is to maximise the probability of reaching a financial target of ~5–10% a year while reducing the risk of catastrophic loss, the trade-off is attractive.

Personally — even expecting lower returns going forward than the backtest — I’d still invest in it, for its diversification, adaptability, and downside protection. But I wouldn’t put 100% of my portfolio into it. We view tactical strategies as one building block on top of a robust strategic core. One simple approach: allocate a portion to Pivot5 and keep the rest in a diversified buy-and-hold core. Here’s 40% Pivot5 / 60% VT:

Growth of $10,000 — strategy alone, 40/60 blend, and VT alone (log scale). A 40/60 blend captures a meaningful share of the downside protection while keeping significant equity exposure, ending well above 100% VT with materially smaller drawdowns. Monthly rebalanced.

Blend metrics approximated from the standalone monthly return series (correlation 0.55), rebalanced monthly.

How has it done recently?

Since the start of 2025, very well. Between 31 December 2024 and 31 May 2026 the strategy returned 42.2% versus 35.4% for VT, outperforming by 6.9 points — a $100,000 investment grew to roughly $142,200 versus $135,400 for the benchmark. What’s interesting is where those returns came from. In 2025, gold did the heavy lifting (+11.6 points), with international developed equities a quiet second engine (+8.3). In 2026 leadership shifted: emerging markets took over (+10.5) and commodities added 4.8 — having detracted slightly in 2025, been removed, then re-entered just before a strong rally. US equities contributed throughout (+2.5) but were never the dominant driver. The lesson: the strategy never needed a particular asset class to outperform — it just needed some assets to be trending, and it rotated accordingly.

Current positioning

As of 31 May 2026, the strategy is fully invested and firmly in risk-on mode:

• 20% Emerging markets (EEM)
• 20% Broad commodities (DBC)
• 20% Gold (GLD)
• 20% US equities (SPY)
• 20% International developed equities (VEA)

There is currently no allocation to bonds, REITs, or cash.

Important information. Educational and informational purposes only — nothing here is investment advice, a personal recommendation, or an offer or solicitation, and it doesn’t consider your individual circumstances. InvestLab is not authorised or regulated by the Financial Conduct Authority. All performance figures are hypothetical and backtested (InvestLab run TAA-monthly-d5ec060407b8, net of 10 bps round-trip costs): they don’t reflect actual trading, and where today’s ETFs didn’t exist, proxy index data was used to extend the history (SPY before 1993-01, IEF/TLT before 2002-07, EEM before 2003-04, VNQ before 2004-09, GLD before 2004-11, DBC before 2006-02, VEA before 2007-07, BNDX before 2013-06, and the VT benchmark before 2008-06). The 35-year window includes a sustained fall in interest rates, a commodity supercycle, and one of the strongest equity eras on record; future conditions may differ. All investing involves risk, including loss of capital, and the strategy can and does lose money. Past performance is not a reliable indicator of future results. The author invests in this strategy or close variants of it. Figures are as of 31 May 2026.