The last two releases were themed batches: momentum in June, the DMS legacy set before that. Themed releases keep the narrative clean, but they also mean every strategy in a batch tends to fail in the same conditions at the same time. This time I worked through the entire unreleased research vault, ranked everything on the numbers our own production backtests generate, and picked for variety rather than cohesion: six strategies that make their money through six different mechanisms. A growth-cycle momentum model, a two-system momentum blend, a channel breakout from a book, a defensive trend switch, a bond timing model, and one fixed allocation.
The bar was simple: every pick clears the live catalog's median Sharpe of 0.97. With this batch, the catalog reaches 88 published strategies. The numbers in this post are the numbers on the strategy pages, nothing massaged. Here is what made the cut and why.
GGCEM
GGCEM (Global Growth Cycle Enhanced Momentum) was developed by Grzegorz Link. The engine is a macro regime filter derived from the OECD Composite Leading Indicator diffusion index, which tracks how many member economies are currently in expansion. When the index signals expansion, the model holds whichever of US or international equities has the stronger 12-month return. When contraction is the reading, it holds whichever of aggregate bonds or T-bills is currently stronger. The signal changes the available universe before the momentum ranking even starts.
The primary variant (Link Original), backtested from 1986, shows 14.1% CAGR, a 1.35 Sharpe, and a -21.0% worst drawdown. That Sharpe is the highest in the batch; the drawdown is real and worth planning for. Three variants ship.
KISS Momentum
KISS Momentum originated from community research shared on Reddit; the author is not named. The design blends two momentum approaches equally: an 8-asset risk-parity system that uses multiple lookback periods, selects the top four assets, and weights by inverse volatility; and a concentrated year-to-date momentum picker that parks in cash during December and January. Neither approach is unusual on its own. Combined, they cover failure modes that neither handles alone.
The combined 50/50 variant, backtested from 1986, shows 10.7% CAGR, a 1.25 Sharpe, and a -22.9% worst drawdown. It fills the two-system blend slot in the cocktail. Four variants ship, including a leveraged one for those who want to stretch the return profile.
Stoken's ACA
Dick Stoken's Active Combined Asset strategy comes from his book Survival of the Fittest for Investors. The structure is three equal-weight opposing pairs (SPY/IEF, GLD/TLT, VNQ/IEF), with the winner of each pair determined by daily Donchian channel breakouts using asymmetric lookbacks. The Dynamic Bond variation adds one rule: when the defensive asset in a pair fails its own channel test, that sleeve moves to cash instead. Turnover averages roughly three trades per year.
Backtested from 1985, the strategy shows 10.4% CAGR, a 1.13 Sharpe, and a -18.9% worst drawdown. The Sharpe is the lowest in the batch, but the mechanism is different from anything else here: a channel breakout rule from a published book, still clearing the selection bar cleanly on its own numbers.
Faber's 12-Month High Switch
Meb Faber's 12-Month High Switch holds each of five risky assets only while it trades within 5% of its 12-month high; weight from any asset that fails that test moves to the defensive side. The Dynamic Bond primary variant applies the same nearness-to-high test to IEF and falls to cash when IEF itself fails. That last detail matters more than it might look. In his post Where is the Yield?!, Meb Faber makes the case that risky bond yield spreads over T-bills are currently thin, which is precisely the environment where a defensive rule that treats cash as a live choice rather than a default parking spot earns its keep.
Primary variant numbers from 1986: 8.4% CAGR, 1.20 Sharpe, -13.6% worst drawdown. This strategy earns its place through the drawdown profile, not the headline return number.
Predicting US Treasury Returns
This strategy implements in ETF form the four-signal bond timing ensemble from Baltussen, Martens, and Penninga's paper Predicting Bond Returns: 70 Years of International Evidence. The four signals are a yield proxy, bond trend, inverted equity returns, and inverted commodity returns, combined into a continuous allocation between IEF and T-bills that ranges from 0% to 100% in IEF depending on the signal composite. The same logic that drives the Faber Dynamic Bond fallback is built more formally into this model's structure: cash is a real alternative to bonds, not a holding position.
From 1988: 4.6% CAGR, 3.9% volatility, 1.19 Sharpe, -10.4% worst drawdown. This is a low-volatility bond sleeve for blending, not a headline return play. Its purpose is to earn a reasonable risk-adjusted return in the environments where holding intermediate Treasuries outright gets uncomfortable.
Desert Portfolio
The Desert Portfolio is a fixed three-fund allocation from the Gyroscopic Investing forum. No signals, no regime filter, no momentum lookback. From 1987: 7.6% CAGR, 1.26 Sharpe, -15.9% worst drawdown. The batch needed one pick that earns its place without any timing mechanism, as a grounded reference for what a well-constructed passive allocation actually produces. This is that pick.
Live July 28, signals ready for the August rebalance
All six strategies go live on July 28, with full backtests and current signals ready before the August rebalance. All six are Pro; the free tier is unchanged.
- GGCEM - a macro regime momentum model that switches between equities and bonds based on OECD expansion signals
- KISS Momentum - a 50/50 blend of a risk-parity momentum system and a seasonal momentum picker
- Stoken's ACA - a channel breakout strategy across three equal-weight opposing asset pairs
- Faber's 12-Month High Switch - a proximity-to-high filter across five risky assets with a cash-capable defensive leg
- Predicting US Treasury Returns - a four-signal bond timing model designed as a low-volatility blend sleeve
- Desert Portfolio - a fixed three-fund passive allocation with no timing rules
Backtests extending before ETF inception use BestFolio's documented proxy fallback chains. All drawdown figures are daily worst drawdowns, consistent with the numbers shown on each strategy page.
Past performance does not guarantee future results. Backtested results are hypothetical and do not represent actual trading.