December 2025: −4.70%, drawdown deepens to −4.9%, no per-instrument breakdown
Live, real-capital, fully systematic.
Live track record → · Methodology →
The numbers
| | This month | YTD | Since inception |
| ------------------------ | ---------- | ------- | --------------- |
| Return | −4.70% | −2.28% | −2.28% |
| Sharpe (ann.) | −3.36 | −0.82 | −0.82 |
| Max drawdown (in window) | −4.72% | −4.97% | −4.97% |
| Hit rate (daily) | 45.5% | 48.7% | 48.7% |
| Drawdown state at MTD | −4.9% from peak | | |
What drove it
Per-instrument attribution is not available for December — daily_attribution.total_pnl_pct was empty for every instrument in this window. The headline number above is the sum of strategy-level daily P&L; the per-instrument breakdown that the dashboard normally surfaces did not reconstruct from upstream reports for this period.
At the strategy level: −4.70% with a daily Sharpe of −3.36 (the worst monthly Sharpe in the captured window). The hit rate dropped from 52.9% in November to 45.5%, and the max intra-month drawdown of −4.72% was not recovered — the system finished December −4.9% from peak, the deepest drawdown to date in this window. The shape of the move (low hit rate combined with a negative monthly mean) is consistent with a regime change against the system’s prevailing positions rather than noise on a still-favourable signal.
How the rules performed
Rule-family attribution is not yet available for December — daily_attribution.rule_contributions was empty for the days in this window. This section will become quantitative once the persistence pipeline has accumulated a full month of data.
What changed
No methodology or operational changes this month.
Current positioning
Annualised portfolio risk: —
Margin usage: —
Looking forward
January begins with the system −4.9% from peak — the deepest drawdown captured in this run. Position-level data for end-December was not persisted, so the January distribution of contributors will be the first read on which trades broke. Two months of data (a +2.54% gain and a −4.70% loss) is far too few to draw any inference about signal degradation; the order matters less than what the next two or three months show. The published methodology and forecast caps did not change in December (What changed is empty); the loss came through the same rules that produced November’s gain.
Past performance does not guarantee future results. Net of estimated transaction costs. Personal capital, not soliciting investors. Full methodology and disclaimers on the live dashboard.