Hi YXI friends,
Here is a quick announcement about how our machine learning model service will evolve in 2026.
DISCLAIMER: This newsletter is strictly educational. Any information or analysis in this note is not an offer to sell or the solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice and nor should it be relied upon to make investment decisions. Any opinions, analyses, or probabilities expressed in this note are those of the author as of the note's date of publication and are subject to change without notice.
Please check out our glossary for explaining the acronyms, financial terminology, and economic data.
1. Machine-Learning Model Update
What’s changed?
The core ML models are unchanged. They continue to combine macro asset signals with individual ticker price features to generate a short-term assessment on the Risk On vs Risk Off status of each asset.
We are introducing a minimum holding-period overlay.
Before this, all of our positions have a minimum holding period of just 1 day, following exactly what each day’s signal says. However, in the future, some assets will have a minimum holding period (in both Risk On and Risk Off) of 2 or 3 days.
This holding period is asset-specific, reflecting different volatility and trend characteristics. They have been backtested with our individual model parameters for optimisation.
What does this mean to you?
The changes take effect on Friday, January 2nd, 2026.
Nothing changes in terms of what you see or act on.
All published Risk-On / Risk-Off signals reflect the final output, so there’s no added complexity or ambiguity.
How often are the models reviewed?
Models are reviewed quarterly and adjusted only when necessary.
The real edge of the system is its ability to protect capital during drawdowns while capturing upside early in new trends. This strength becomes clear over a complete price cycle.
This means it’s important not to overreact or overoptimise the models based on recent price action, as this increases the risk of overfitting. The models themselves already account for the ongoing changes in the market dynamics.
2. Current Machine Learning Signal Offerings
We offer three sets of machine-learning signals.
Macro Assets and Megacaps
The following are offered as part of this flagship service from January 2nd. IWM and SOXX are new additions.
Macro: SPY, QQQ, IWM, SOXX, BTC, TLT, UUP
Equities: AAPL, AMZN, GOOGL, META, MSFT, NVDA, TSLA, PLTR
Crypto
The following are offered as the Crypto ML Signal package, hosted on Cestrian Capital Research’s Slack channel.
BTC, ETH, SOL, XRP, COIN, HOOD, MSTR.
Once we have 1 year of trading data, we will add BMNR and SBET as well.
Commodities
The following are offered as the Commodities ML Signal package, hosted on Cestrian Capital Research’s Slack channel.
GLD, GDX, SLV, SIL, USO, PPLT, PALL, CPER, URA, BCI
URA and BCI are both new additions for 2026.
Thank You
I wish you all a lovely New Year.
2026 will undoubtedly be a challenging year in the markets, but it will also offer many opportunities.
My biggest goal for 2026 is to help you spot rotational opportunities across stocks, crypto, bonds, and commodities to make money in all weather conditions.
Thank you!
Yimin