How different ways of choosing trend following programs have performed over time
Data through July 2026. Edition date 27 September 2026.
The study uses reported returns for programs classified in NilssonHedge as Type CTA, Style Systematic and Strategy Trend Following. We compare 29 versions of allocation methods in three settings: a $10 million investor choosing ten managers, a $100 million investor choosing ten, and a $100 million investor choosing five. The capital figure is the investor’s total portfolio, divided equally among the programs selected; reported program minimums must fit each starting allocation.
At the end of each December, every method selects the programs it will hold for the next year. In January, it allocates an equal amount to each selected program. It holds those positions through December, then makes a new selection for the following year.
This edition first ranks every allocation option by growth since January 2000, then by return relative to its deepest fall (MAR). We then examine July’s five leaders and the five leaders on each long-run measure, CAGR and MAR. SG Trend and SG CTA are the index benchmarks throughout.
July’s leading method gained 1.15%, while SG Trend fell 1.12% and SG CTA fell 1.26%. A single month’s lead does not settle which allocation method has worked best over the full period.
How to read this report
Complete ranking by CAGR, January 2000 to July 2026
All 32 allocation options and the two index benchmarks, ranked by the headline measure over January 2000 to July 2026 (2026 is a partial year). The index benchmarks sit at the positions their own figures earn and are highlighted. Equal-weight options are variable-size references and are not eligible to be a monthly winner.
Complete ranking by MAR, January 2000 to July 2026
All 32 allocation options and the two index benchmarks, ranked by the headline measure over January 2000 to July 2026 (2026 is a partial year). The index benchmarks sit at the positions their own figures earn and are highlighted. Equal-weight options are variable-size references and are not eligible to be a monthly winner.
Top 5 by CAGR from six starting dates, each ending July 2026
Each row restarts the clock in January of the year shown, using each method’s same published monthly path rebased to 100; no selection is rerun. Rows rank on that period’s unrounded figure. The January 2025 row covers only 19 months and is short and partial: its annualised CAGR and MAR can move sharply with a few monthly results. SG Trend and SG CTA are shown beside each row as benchmarks, not as candidates.
Top 5 by MAR from six starting dates, each ending July 2026
Built the same way as the CAGR grid on the previous page, but ranked on each period’s unrounded MAR. SG Trend and SG CTA are shown beside each row as benchmarks, not as candidates.
Explore The Allocation Ledger
Growth of 100 from January 2000, month by month. Choose methods by Fund #, or use a preset; step or play through time; SG Trend and SG CTA are the index benchmarks.
Select at least one Fund # or index benchmark to draw the chart.
| Shown | Growth of 100 | CAGR since 2000 | MAR since 2000 |
|---|
Investor settings and minimum investments
The $10 million and $100 million figures describe the investor’s total capital. Each method checks whether a program’s reported minimum investment fits the amount initially allocated to it:
A reported minimum above that amount excludes the program. When NilssonHedge does not report a minimum, the rule treats it as affordable. The variable-size equal-weight reference checks capital divided by its chosen number of holdings. These minimums come from a later database snapshot, without historical dates or confirmation that the program was open to investment. The screen is a consistent simulation rule, not proof that each allocation could actually have been placed at the time.
Top 5 for July 2026
The five allocation contenders with the highest return in July 2026, with their long-run figures from January 2000. A strong month is descriptive only: it never changes a method’s holdings before the next December review. Each method’s rule and every program it held in July 2026 follow the table.
Top 5 since inception by CAGR
Top 5 since inception by MAR
Three different winners
This edition has three different leaders. F32 (20-year MAR ensemble, 5 programs, $100m) led July 2026 with a return of 1.15%. Over the full period, however, its CAGR of 4.05% and MAR of 0.18 trail both SG benchmarks (SG Trend 5.56% and 0.27; SG CTA 4.39% and 0.27). F22 (3-year CAGR ranking, 5 programs, $100m) grew fastest, with a CAGR of 12.31%, but its deepest fall was 28.7%. F23 (3-year MAR ranking, 5 programs, $100m) has the highest MAR, 0.63: a CAGR of 9.57% with a deepest fall of 15.1%. A month’s leader shows what worked recently, the CAGR leader shows which method grew most, and the MAR leader shows which grew most for the size of its worst decline. The report follows all three because they measure different things.
The annual decision
On 31 December each method chooses its group using only the history available at that date. On 1 January it invests equal amounts in each chosen program. The holdings are not rebalanced during the year, so each program’s share drifts with its returns. A program that does not report in a month is held flat for that month under the reporting convention; that does not mean it has stopped. On the next 31 December the method reviews and chooses again.
Glossary
CAGR. Compound annual growth rate: the steady yearly return that would turn the starting value into the ending value.
Deepest fall. The largest percentage decline from a previous high to a later low (maximum drawdown).
MAR. CAGR divided by the deepest fall. Higher means more growth for each unit of the worst decline.
Growth of 100. What 100 invested in January 2000 would be worth, compounding monthly (VAMI). The charts use a logarithmic vertical scale, so equal steps show equal percentage changes.
Fund #. A permanent identifier for each simulated allocation option (F01 to F32). It is not an actual fund.
Ranking. Scores each program on its own history and takes the top programs, one per manager.
Ensemble. Searches whole equal-weight groups and keeps the group whose combined history scores best. Large searches are not always proven to find the single best group; see the technical appendix.
Long-record pool and early-year proxy. Some methods use only programs with 20 years of history. Before enough such programs existed, they used the longest history that still left enough managers, a rule frozen in advance.
Reporting convention. A held program with no return for a month keeps its value for that month (a zero return). Nothing is redistributed, and a late return goes to a revision queue rather than changing a published month.
Index benchmark. SG Trend and SG CTA, published indices of trend following and managed futures programs. They are references, never monthly winners.
Variable-size reference. Equal weight across every affordable manager. Its size varies, so it is not eligible to be a monthly winner.
Notes
- Each edition adds the latest complete month. Earlier published results are retained.
- Returns are as reported by each program and index, which generally means after the fees each one charges. This report deducts no further fees or costs, and programs and indices may not calculate their returns on the same fee basis.
- Many methods were examined, so a leader identified after the fact carries a selection advantage. Past results do not predict future returns.
- Edition built from ledger vintage V2026-07-founding, revision r3; engine SHA-256 1d1264085811; cutoff configuration SHA-256 231150f6d180; Fund # mapping SHA-256 fab8e49870c1.
Richard Brennan writes on systematic trading, complex adaptive markets, and the philosophical foundations of trend following at atstradingsolutions.com. His books include The Fractals of Finance, Complex Adaptive Markets, Carved by Impossibility and The Aussie Turtles Trend Following Guide.
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.