A Major YXI Model Strategy Upgrade (April 13, 2026)

DISCLAIMER: This newsletter is intended for educational purposes only. Any information or analysis in this note does not constitute an offer to sell or a solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice, 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.

Hi YXI friends,

I have been working on a major upgrade to our signal models.

I want to walk you through how the system works and introduce two new optional strategy tiers: Smart Leverage and Smart Short.

The models are built on a simple premise: asset prices are driven by a combination of macro forces and asset-specific dynamics, and the relative importance of those forces shifts over time.

Each model reads two categories of input.

The models are trained on multiple years of data, but they are not static. They recalibrate daily on a rolling basis, adapting to current conditions. A factor that mattered enormously during the 2022 rate-hiking cycle may carry less weight today, and the models adjust accordingly. This makes them dynamic rather than backwards-looking.

Each model produces an internal conviction gauge per ticker per day. When that gauge crosses above a certain threshold, the signal flips to Risk On. When it drops below a lower threshold, the signal flips to Risk Off.

There is also a minimum holding period to prevent whipsaw: once a position is entered, it must be held for a set number of days before a new signal can change it. All of these parameters are calibrated individually per ticker, because SPY and GLD behave differently, and the models should reflect that.

The output, at its most basic level, is a binary signal: Risk On (long) or Risk Off (flat). No leverage, no shorting, no options. Clean and simple.

Signals are delivered daily before the US market open. Each morning, you receive:

Signal Snapshot. The current signal for every ticker in the suite, with position changes clearly flagged. This is the actionable output.

Signal Commentary. Every day, the commentary explains what the top drivers are behind the signals. Which macro factors are dominating? Why did a signal flip, or why is it holding? This is not a black box. You see the reasoning.

Over time, reading the commentary genuinely improves your understanding of cross-asset dynamics and regime shifts, regardless of whether you follow the signals mechanically. Many subscribers have told us the commentary alone is worth the subscription.

Position History. A running log of every signal change: entry dates, exit dates, and the outcome for each position. Full transparency.

We track performance every Monday: frequent enough to stay on top of the signal calibration, but not a daily whipsaw distraction.

This is the upgrade.

The baseline model operates in two states: flat or long. That has served us well, and it remains the default. But the model's conviction gauge is not binary. It has a range. When conviction is very high, the model is not just bullish but strongly bullish. When conviction is very low, the model is not just cautious but actively bearish.

The new strategy tiers exploit that gradient.

Smart Leverage adds a third state: 2x long.

Smart Leverage + Short adds a fourth state: short.

The thresholds for leverage and shorting are optimised individually per ticker. Every ticker gets its own calibration.

Here is the snapshot you can expect going forward:

To be clear - this is just a mockup showing the full upgrade, not the real data.

We have removed the “streaks” and “10-day average” metrics as our history position table already accounts for them. Instead, we have added columns to show when “Leverage” or “Short” strategies are active and what the model does in the “Notes” column, to avoid any confusion.

The historical tables will have new colours to indicate leveraged and short days:

1* means leveraged long, and 0* means short. (Again, this screenshot is a mockup.)

In backtesting, both tiers have improved Sharpe ratios across the majority of tickers without materially increasing maximum drawdowns.

Here is an example of the frequencies we would expect to see shorts and leverage states being active, using SPY as an example.

Our backtests have assumed leveraging and shorting the underlying spot ETFs with borrowing costs accounted.

However, I will begin looking for the corresponding leveraged and inverse ETFs for our assets to achieve the most appropriate approximation.

Given the number of models, please allow me some time to complete the full update (ETA: 1 week).

The new strategy tiers are entirely optional.

The baseline Risk On / Risk Off signal remains the core product and is unchanged. If you are following the baseline system and it is working for you, there is no need to change anything.

The leverage and short tiers are for subscribers who want to explore more aggressive positioning using the same underlying model. They layer on top of the baseline. They do not replace it.

The model’s new strategies could frequently disagree with your own analysis. As we have done along this journey, we collect live results, watch the model's reaction vs our expectations, and see how the real markets unfold.

Every quarter, we recalibrate and look for potential optimisation while trying to minimise over-fitting bias.

We will track all three strategy tiers separately and transparently. Every Monday, the performance update will include metrics for baseline, +Leverage, and +Leverage & Short: cumulative return, Sharpe ratio, and maximum drawdown. You will see exactly what each tier adds or costs relative to the baseline.

All P&L assumes next-open execution as before.

Baseline only. Follow the binary signals as you always have. This remains the simplest and most battle-tested approach.

Baseline + Leverage. When the model has high conviction and volatility is low, one sizes up. When conviction drops, one scales back. Long-only throughout.

Full system. Baseline + Leverage + Short. This is the most aggressive option. It is for experienced traders who are comfortable with short selling and its associated risks.

You do not have to choose one tier across the board. You can follow baseline on some tickers and leverage on others. Signals for all three tiers are delivered simultaneously.

Strong backtest results are never a guarantee of live performance. We will monitor how the new strategies perform in live markets across different conditions before drawing conclusions. If the results are not convincing, we stay with the baseline. Prudence over excitement.

Separately, our models show their real edge over a full market cycle, not just the good parts. They can lag buy-and-hold in a strong uptrend because the model occasionally steps aside. But over a full cycle, the capital saved during drawdowns compounds into higher returns. We have just lived through a complete crypto cycle, and the results bear this out.

I will keep you updated. As always: clarity, not noise, and a real edge in the market.

Thank you,

Yimin