Journal

Notes from the desk

Research notes on how we build, size, and test the portfolio.

• The Alpha Break-Even Cost: Where Trading Edges Go to Die A momentum strategy returned 9.17% at zero costs. At 35 bps per trade, its edge vanished. At 50 bps, 49% of the promised wealth was destroyed.
• Portfolio Beta Isn't a Constant: A 20-Year Test of the 60/40 Risk Narrative A 60/40 portfolio's beta to the S&P 500 ranged from 0.30 to 0.79 over 20 years. The stock-bond correlation regime, not the allocation, drives the swing.
• Causal Structure, Not Just Parameters: Why Regime Switching Matters for Options Market regimes are different causal structures, not parameters. Evidence is mixed on accuracy but agrees: markets change structure.
• The Equity Risk Premium Is Not a Number — It Is a Distribution The realized ERP averaged 12.15% over rolling 10-year windows from 2007-2026 — nearly triple the 4.33% implied estimate. But 17% of years went negative.
• Beyond Sharpe: 6 Metrics That Reveal a Strategy’s True Risk Three strategies, same Sharpe ratio, wildly different risk. Six metrics that reveal what Sharpe misses.
• Chasing the Ghost: Why AI on Price Indicators Finds Overfitting, Not Edge Most trading indicators are the same signal in different formulas. Stacking them with AI adds overfitting, not edge. What grounds a system is a mechanism you can explain.
• How Much Should You Bet? A Backtest of Position Sizing Methods Four position sizing methods backtested on 15 years of SPY data. Buy-and-hold, trend filter, volatility-targeting, and the combination — with real numbers.
• The Broken Slot Machine: Why Alpha Erodes, and What Stays Alpha is real, but it erodes. Every edge ever found has been slowly arbitraged away by the market — a self-equilibrating machine that fixes its own glitches. Yet some things persist.
• When Data Stops Being the Edge: Feature Engineering, LLMs, and What Ordinary Investors Can Learn from Two Sigma Ben Wellington of Two Sigma explains why the quant edge moved from data to feature engineering. LLMs change the game — but discipline, not data, is the new moat.
• Three-Fund vs Golden Butterfly vs HFEA: What 15 Years of Backtesting Actually Shows Three popular portfolios, 15 years of real backtest data, one uncomfortable truth: they all harvest the same equity risk premium. Leverage scales it. Complexity doesn't change it. True diversification requires fundamentally different risk premia.
• Why volatility is the only variable you can control The one portfolio variable you can actually control. A real SPY backtest shows why scaling your position to a volatility target beats letting the market decide your risk for you.
• Just keep buying The simplest idea in investing — and the one that quietly builds the most wealth. What Nick Maggiulli's book gets right, in numbers.
• An introduction to harvesting risk premia Why most so-called alpha was never alpha — and how ordinary investors can capture the risks that actually pay, while skipping the ones that don't.
• Diversification: the only free lunch Harry Markowitz called it the only free lunch in investing. A tiny example shows why it's nearly magic — and why real diversification is about understood drivers, not a correlation number.