Rotates between growth, neutral and defensive ETF portfolios by reading the direction of US real interest rates, with a volatility circuit breaker on top.
Hypothetical results including 5 bp trading cost per side. Designed and tuned on 2003–2024 data; 2025 onward is forward-test data that was never used for tuning. Data to 2026-09-23. Past performance does not guarantee future results.
Sharpe and Sortino use a 0% risk-free rate and daily returns. Win rate is the share of calendar months with a positive return.
How far the portfolio fell from its previous high, and how it did month by month.
A great backtest proves little on its own. These tests ask whether the result could be luck, whether it holds up across market conditions, and whether the record is long enough to trust.
Combines the three tests below into one 0–100 number: how likely the historical performance is to continue rather than being a product of chance.
Preliminary score. The scoring method will be published in How it works.
Sample size, fat tails and streaks can all make a lucky strategy look skilled. We test for that directly.
Annualised return over random 5-year stretches stitched together from the strategy’s own daily history in 1-month blocks. It shows the spread of outcomes the record supports, not a forecast.
Built to hold up when stocks fall, not just when they rise. Stocks, gold, long bonds and cash-like Treasuries are all in its toolkit.
The rolling line shows the risk-adjusted return of every 3-year window since 2003. A robust strategy stays above zero in all of them.
More than two decades, including the 2008 financial crisis, the 2020 crash and the 2022 rate shock.
Six liquid US-listed ETFs. The strategy is long-only, fully invested and unleveraged, and moves between them as conditions change.
Average portfolio mix in each month. Follows the time range above.
Mix of the combined reference portfolio, averaged over each month. IEF stands in for GLD before Nov 2004 and for BIL before May 2007, the dates those ETFs launched.
Every subscriber gets the same signal at the same time. Signals shown here were produced by running the strategy code on historical data (backtest). They were not sent at the time.
Delivered by email whenever the strategy trades. Unsubscribe any time.
Educational information only, not investment advice. oxrun does not execute trades or hold funds, and is not a registered investment adviser, broker-dealer or CTA. Backtested and forward-tested results are hypothetical, include modelled trading costs and do not reflect actual trading. Past performance does not guarantee future results.