$100 invested in January 2000 with F22, a simple rule for choosing five trend followers each year, would have grown to $2,347 by August 2026 in our simulation. The same $100 in the SG Trend Index grew to $434, in the SG CTA Index to $322, and in the S&P 500 Total Return Index to $883.
This is the first Allocation Spotlight, a monthly companion to The Allocation Ledger. Each month we take one of the 32 allocation rules the Ledger tracks and explain how it works, what it holds today and what its record shows. We start with F22 because it has grown fastest of the 32 since 2000. We can say that only because we can see all 32 records, so the result is best read as a question: what does a simple, mechanical rule reveal about the benchmarks investors use for trend following?
Three measures carry the story. CAGR, the compound annual growth rate, is the steady yearly rate that turns the starting value into the ending one. Drawdown is the fall from a previous peak. MAR divides CAGR by the deepest drawdown, so it shows how much growth an investor earned for the worst decline they had to sit through. F22’s CAGR was 12.56%, more than double SG Trend’s 5.66% and nearly three times SG CTA’s 4.49%. It also beat the S&P 500 TR on growth (8.51%) and on MAR, 0.44 against 0.17.
F22’s results are simulated from the reported returns of real programs under a fixed rule; they are not the record of an actual fund. Its returns are after Manager Fees, meaning the underlying managers’ reported management and performance fees, as are the SG indices, and before any additional allocation, trading or implementation costs. Much of its growth came from a small number of exceptional months, and that pattern is the centre of this article.
What the SG indices measure
The SG indices measure a defined group of large managers, not the full range of what trend following can deliver. Since January 2013 the SG Trend Index has held the ten largest trend following CTAs that meet Société Générale’s criteria on strategy, diversification, correlation, reporting and openness to new investment, equal weighted and reconstituted each year. The SG CTA Index applies a similar approach to the 20 largest CTAs, trend followers or not.
That makes them useful records of what the industry’s largest programs have delivered. It also makes them one particular, size-based slice of the field. F22 explores a different question: what a simple selection process would have produced from the wider set of programs in the database.
The Ledger shows where the indices sit within that wider set. Plot every allocation option it tracks by CAGR and by MAR, and both SG indices land in the bottom-left quadrant: below the median option on growth and below the median option on MAR. These are simple, mechanical rules, yet most of them beat both indices on both measures.
The Allocation Ledger, August 2026 edition: all 32 allocation options (F01 to F32) and both SG indices, January 2000 to August 2026.
Even the plainest reference in the Ledger, holding every affordable trend follower in equal weight, roughly matched SG Trend at 5.62% to 5.75% a year. These results challenge treating the SG indices as the full measure of what trend following can offer. They do not explain why the gap exists. Differences in selection, market exposure and implementation could all contribute, and one plausible factor, that the largest firms manage for capacity and moderate volatility, is a hypothesis this comparison cannot isolate.
How F22 is built
F22 is the 3-year CAGR ranking with five programs and $100 million of capital. It makes one decision a year, ranking programs using returns dated through the preceding December.
- Start with the trend followers. The candidates are every program in the NilssonHedge database classified as Type CTA, Style Systematic and Strategy Trend Following.
- Keep what the investor can buy. F22 is tested for a $100 million investor, which allows up to $20 million per holding, so a program qualifies only if its reported minimum investment is $20 million or less.
- Require a real record. A program needs at least three years of unbroken monthly returns at the review date.
- Rank on growth. On 31 December, each eligible program is ranked on its own compound annual growth rate over the previous three years.
- Take the top five, one per manager. If a manager has two programs near the top, only its higher ranked program is kept.
- Invest equally and leave it alone. On 1 January each holding receives 20%. There is no rebalancing during the year, so each holding’s share drifts with its own returns.
- Next 31 December the rule ranks the whole field again and may replace any or all five.
There is no optimisation, no forecast and no judgement call. The rule does not search combinations of programs; it ranks them one at a time.
The 2026 selection
The review on 31 December 2025 ranked programs on their returns from January 2023 to December 2025. These five each received 20% on 1 January 2026, and their shares have drifted with their returns since:
The group gained 7.15% in August and 22.63% in 2026 to the end of August. Over the same eight months SG Trend gained 11.07% and the S&P 500 TR 13.14%. These holdings stay in place until the next review on 31 December 2026.
Growth of $100 since January 2000
F22, SG Trend and SG CTA: The Allocation Ledger, August 2026 edition. S&P 500 TR: ATS Master Performance Template. Quarter-end values plus August 2026.
The gap opened early and kept widening. F22 compounded faster than the S&P 500 TR and far faster than either SG index, and it did so with a shallower worst decline than the equity market.
January 2000 to August 2026, 320 months. MAR is CAGR divided by the deepest fall. F22 and the SG indices are after Manager Fees; F22 is before any additional allocation or implementation costs. The S&P 500 TR is an index with no fees deducted.
Convexity: where F22's growth comes from
F22’s long-run growth depended heavily on a small number of exceptional months. Three related properties are worth separating here. Right-tail dependence means a few large gains do most of the compounding. Positive skew means monthly returns lean toward large gains rather than large losses. Convexity describes an upward-bending relationship between returns and a specified market move. A smile pattern can indicate gains in both strongly rising and strongly falling markets. F22’s record shows the first strongly, the second mildly and the third as a historical pattern.
Right-tail dependence. F22’s best 32 months, one month in ten, account for all of its compounded growth since 2000. The other 288 months together came to a 5.6% loss. SG Trend’s best months were smaller, and its other 288 months together cost it 55%, so its gains did far less to offset the months in between.
Skew. F22’s average winning month (+4.40%) was 1.26 times the size of its average losing month (−3.50%). For SG Trend the ratio was 1.13, and for the S&P 500 TR it was 0.86. F22’s monthly returns lean slightly to the upside (skew +0.22), while the S&P 500 TR’s lean to the downside (skew −0.47). These are modest asymmetries.
A historical smile. The chart below sorts the 320 months into five groups of 64, from the S&P 500 TR’s worst months to its best, and shows each series’ average return in each group.
F22 and SG Trend: The Allocation Ledger, August 2026 edition. S&P 500 TR: ATS Master Performance Template. January 2000 to August 2026.
Over this period F22 was positive at both ends: it gained most in the equity market’s worst months and still gained in its best. It was not a perfect smile, since in the second-worst group F22 lost slightly more than the market. The pattern describes the last 26 years; it does not guarantee protection in future crises.
What the rule contributes
F22 does not forecast trends, and its record does not show that ranking on recent growth reliably selects managers with convex returns. What its structure does is shape how exceptional results flow through the portfolio.
- It ranks on recent compounded growth. With no volatility penalty, managers whose last three years include large gains rank highly. High growth can come from several return patterns and levels of exposure, so this is a tilt toward recent strength, not a guarantee of convexity.
- It concentrates. With five holdings, one manager catching a major trend moves the whole portfolio. So does one manager doing badly.
- It lets successful holdings grow within the year. Each holding starts at 20% and is not rebalanced, so weight drifts toward the managers that are working. The rule does not exit a losing holding during the year; it lets that holding’s share shrink.
- It renews each year. Managers that fall out of the top five are replaced. That keeps the portfolio pointed at recent strength, although a manager dropping out of the ranking does not prove its edge has gone.
These features have a cost. F22’s deepest fall was 28.7%, against 20.7% for SG Trend, and it lost money in 2005, 2009, 2011, 2013 and 2016. The fair reading of the evidence is that F22’s historical growth depended on exceptional months, and its rules allowed the holdings producing them to become larger during the year. Whether that pattern persists is what the Ledger now tests month by month.
F22 against the S&P 500 TR
F22 beat the S&P 500 TR on growth, 12.56% a year against 8.51%, and on risk-adjusted growth, with a MAR of 0.44 against 0.17. MAR divides growth by the deepest fall, so it asks how much an investor earned for the worst loss they had to sit through. The S&P 500 TR’s deepest fall was 50.9%, from October 2007 to February 2009. F22’s was 28.7%, from February 2004 to April 2005, and it was back at a new high by September 2007.
F22: The Allocation Ledger, August 2026 edition, after Manager Fees. S&P 500 TR: ATS Master Performance Template. Values every second month.
F22’s falls came more often but went less deep. It had three declines of more than 20%, bottoming in April 2005, August 2013 and October 2024, and was back at a new high within two and a half years each time. The S&P 500 TR took until March 2012 to regain its October 2007 high.
F22 did not win by beating equities most of the time. It finished ahead of the S&P 500 TR in only 12 of the 26 full calendar years from 2000 to 2025. It won on the size of its gains when it mattered: in every one of the six years the S&P 500 TR fell, F22 rose.
SG Trend did better than F22 in 2022, and matched it in the early-2000s bear market. The difference over 26 years lies mostly in the years between crises, when F22 kept compounding and SG Trend’s growth stalled for long stretches. Across the full period, F22’s monthly returns had a correlation of 0.00 with the S&P 500 TR. That means no observed linear relationship in this sample, which made F22 a source of growth largely separate from equity market moves over the period.
How the data was handled
Every result here comes from the same fixed process the Ledger applies to all 32 rules.
- Each simulated selection excludes returns dated after the review date. The holdings each review picks are recorded and never changed afterwards. Each rule’s definition is frozen, applied the same way at every review, and never retuned with later data.
- Reporting lag has been tested. Managers do not all report December returns by 31 December. We re-ran the rule with returns to 31 December starting the allocation on 1 February rather than 1 January, and the effect on the results was immaterial.
- Closed programs stay in. The candidates at each review include programs that later stopped reporting. NilssonHedge keeps their records, so they are not quietly dropped from history.
- Missing months are held flat. If a held program stops reporting during the year, its stake is assumed to earn nothing until the next review. That is an assumption: the real outcome could have included later gains, losses or redemption delays.
- Returns are after Manager Fees, the underlying managers’ reported management and performance fees, as are SG Trend and SG CTA. They are before any additional allocation, trading or implementation costs. The S&P 500 TR is an index with no fees deducted.
- Published months are locked. The Ledger is append-only. If NilssonHedge later revises an earlier month, the change is logged and queued, not written over the published record.
Some limits remain. A return dated before a review does not prove the record was in the database at that time, so retaining closed programs reduces survivorship bias but cannot rule out backfill or the bias of managers choosing whether to report. Minimum investments come from a recent snapshot of the database and are applied to all years, and a program with no reported minimum is treated as affordable. And as noted at the start, F22 is featured because it has the best record of the 32 rules; that choice is made with hindsight even though the rule’s own selections are not. This is why the Ledger tracks every rule forward and reports formal tests of luck and search in its technical appendix.
Who can actually invest this way
Unfortunately, the benchmark approaches are far easier to buy than a rule like F22. Several listed funds offer exposure that resembles the SG indices for around 1% a year in fund expenses. None offers F22.
Fund | Exposure | How it works | Annual expenses |
iMGP DBi Managed Futures Strategy ETF (DBMF), NYSE Arca | Resembles SG CTA | An actively managed replication strategy. It estimates the aggregate exposures of a pool of large CTAs from their trailing 60-day performance and expresses them through liquid futures. It aims to simulate their performance, not to reproduce their positions. | 0.85% |
iMGP DBi Managed Futures Fund R USD UCITS ETF (ISIN LU2951555585), Euronext Paris and London | Resembles SG CTA | Broadly the same replication approach, in a separate European fund with its own implementation and regulatory constraints. | 0.75% product cost |
Return Stacked U.S. Stocks & Managed Futures ETF (RSST), Cboe | US equities plus an SG Trend-style sleeve | About $100 of US large-cap equities plus about $100 of managed futures for every $100 invested. The managed futures sleeve blends 70% bottom-up trend signals with 30% top-down replication of SG Trend. | 0.99% |
Return Stacked Bonds & Managed Futures ETF (RSBT), Cboe | US bonds plus the same sleeve | About $100 of US bonds plus about $100 of the same managed futures sleeve. | 1.01% |
KraneShares Mount Lucas Managed Futures Index Strategy ETF (KMLM) | Its own trend index | Follows its own index of commodity, currency and government bond futures rather than replicating the SG indices. | 0.90% |
The Return Stacked funds need a plain-English warning. For every $100 invested, they seek roughly $100 of stock or bond exposure alongside roughly $100 of managed futures exposure, using futures to create the extra exposure. Both parts drive the result, and the leverage can magnify losses. Neither fund is a standalone investment in SG Trend, and we found no ETF that is. KMLM is a reminder that accessible trend following is a wider category than SG replication, with each fund following its own method.
F22 was tested for a $100 million investor, allowing up to $20 million per holding. Whether a smaller investor could reproduce its selections depends on the actual minimums and access conditions of the programs selected. On subscription minimums alone, five equal holdings need five times the largest minimum among the five; investor eligibility, manager capacity and share class also matter. Applying F22’s affordability screen at a smaller level of capital could also produce a different selection.
That gap is an opportunity for product design. A vehicle could pool smaller investors’ capital to hold a rule-based selection through a managed account platform. Or it could apply the replication methods used for the SG indices to a selected group of managers, simulating that group’s combined performance with liquid futures. Either route faces real hurdles: holdings can change every December, a concentrated group of five managers may be harder to replicate if its returns depend on distinctive positions that liquid futures cannot closely reproduce, and some managers limit who can access their programs. Until such vehicles exist, rule-based selections like F22 will remain difficult for most investors to implement.
Expenses are each fund’s published annual operating expenses (for the UCITS ETF, the product cost published by Deutsche Börse). They are not complete implementation costs. Sources: DBMF prospectus, April 2026, and iMGP DBMF page; Deutsche Börse product page for ISIN LU2951555585; Return Stacked RSST page and the issuer’s RSBT page; KraneShares KMLM page. Checked October 2026.
Build your own comparison
Readers who want to test other rules can do so in the current edition of The Allocation Ledger. Its interactive exhibit holds every rule, F01 to F32, alongside SG Trend and SG CTA. Pick any combination by Fund number and the chart redraws their growth from January 2000.
Three preset buttons give quick starting points: the month’s top five, the top five by CAGR since 2000 and the top five by MAR since 2000. Try setting F22 beside F23, the rule that ranks on MAR instead of CAGR, to see how much growth a smoother path gives up. Or set the 3-year rules against the 20-year rules to test whether a longer record makes a better guide.
Every rule follows the same annual clock, the same data and the same reporting conventions, so any comparison you build is like for like.
What comes next
F22’s next review is on 31 December 2026, when it will rank the field on returns from January 2024 to December 2026 and choose its holdings for 2027. We will report the new selection in the January edition of the Ledger. Next month’s Spotlight turns to a different rule.
All figures run from January 2000 to August 2026; 2026 is a partial year. F22’s results are simulated from programs’ reported returns under a fixed rule; they are not the record of an actual fund and do not include allocator fees, trading costs or taxes. Past performance does not guarantee future results. This article is for information only and is not investment advice. Source data: NilssonHedge program returns; SG Trend, SG CTA and S&P 500 TR monthly returns from ATS Trading Solutions’ performance records.
Richard Brennan writes on systematic trading, complex adaptive markets, and the philosophical foundations of trend following at atstradingsolutions.com. His published books include The Fractals of Finance, Complex Adaptive Markets, Carved by Impossibility, and The Aussie Turtles Trend Following Guide, co-authored with Adam Havryliv. His forthcoming book, The Trend Following Manifesto, co-authored with Niels Kaastrup-Larsen, is coming soon.
You do not need to predict the future. You need a process that keeps you present for the rare moves that matter.
The Trend Following Manifesto: Building Robust Portfolios for an Uncertain World
Coming soon to Amazon in paperback and hardcover. Join the book launch list to receive the confirmed launch date, details of a special first-week discount, and personal emails about the story behind the book.
Want to explore why structure exists at all?
Carved by Impossibility: What Remains When Everything Else Is Eliminated
The book explores the architecture of constraint, emergence, and reality itself, and what it means for how we understand markets, life, and the universe.
Available now on Amazon in paperback, hardcover, and Kindle.
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.