Paper: Conformal intervals as Kelly position sizing yield 28.5% annualized net growth
PtrPomorski · x · 2026-08-18
A new arXiv paper, "Conformal Kelly," uses a 75% conformal prediction interval as the scale in fractional Kelly sizing: shrink positions as the interval widens, grow them as it narrows. On a six-year development window (2016-2021) with trading costs and strict leverage caps, it compounds at 28.5% annualized net log growth, Sharpe 1.34, max drawdown 27.7%, vs 15.9% for the S&P 500 and 21-22% for passive portfolios at the same leverage.
Counterintuitively, every tweak that adapts the interval faster to market conditions costs 0.7-5.3 points of annual growth; the simplest method (slow, unweighted, per-asset rolling quantiles) wins — when an interval sizes a position, width stability beats local sharpness. A downside-miss risk control cuts max drawdown to 20.3% while raising Sharpe, beating all 40 placebo timings (p=1/41).
More from Research
- Surgical AI training requires human-defined ontology and verified data — ddonoho · 2026-08-18
- Developer proposes mixture-of-attention architecture with centroid clustering for token routing — InfamousTrouble7993 · 2026-08-18
- Model time perception may be fuzzy interpolation heuristics — kalomaze · 2026-08-18
- Speculation on RL deployment speed affecting model generation — kalomaze · 2026-08-18
- ByteDance Paper: Are AI Agents Actually Controllable? — rohanpaul_ai · 2026-08-18
- Optimizer Geometry Changes Scaling Laws: Muon Outperforms AdamW in Hard-Rank Growth — YouJiacheng · 2026-08-18