In Hindsight 09/29/2025
Welcome to a new series where I revisit previously shared QuantMage spells and check how they behaved out of sample. Think of it as post‑sharing care—a little forward testing to keep us honest. It’s equal parts fun and humbling.
🏆 Winners Beyond the Backtest
Inflation-signaled Gold/Treasury
This one traveled well out of the lab, delivering an uncorrelated return over a six‑month out‑of‑sample window. Half a year is not proof of persistent edge, but it’s a good first chapter.
Yen Carry Trade Unwind
Another solid performer. It held up particularly well during the Liberation Day drop, which speaks to its defensive character during sudden regime shifts.
BDRY Signal
This dodged much of the Liberation Day air‑pocket, but it missed part of the early rebound. Net‑net, performance was roughly in line with the benchmark over seven months, which is respectable.
Credit Rotation via Utilities
Another spell that sidestepped the drop. Absolute return is modest over six months, but that’s by design—low‑risk / low‑return—and it behaved exactly like that.
🕰️ Too Early to Tell
Catching Crypto Trends
The Bitcoin spell has gone mostly sideways over five months. Given muted BTC action in that window, I’m inclined to give it more time before judging the signal.
🪦 The Duds (So Far)
Monthly Seasonality
This one struggled out of the gate. If you look at the first chart above, you can see the trend deviation right away. Without a strong causal story for the month‑by‑month picks, the in‑sample “win” looks more like data mining than durable edge.
⚠️ Reminder: This isn’t investment advice. Just one mage’s systematic scribbles. Always do your own research (DYOR), and consider talking to a qualified advisor before investing.
Conclusion
Rearview mirrors aren’t windshields. Past performance helps with context, not certainty.
Signal pain is ambiguous. It’s hard to tell a rough patch from a broken edge; your thesis‑level conviction is what gets you through the former.
Don’t rent conviction. No pundit—no matter how data‑driven—can hand you belief. Kick the tires yourself and own the decision.
impressive! I mean not only the performance as such, but also the transparency of the process.
Your High Beta vs Low Volatility strategy also performed very well; I guess you didn't include it here because it hasn't been out of sample for six months? Hence, with luck, your next Hindsight piece will have another gem.
For the baltic dry you could blend it with defensive first or another defensive asset allocation during the risk off period