Risk Management After the AI Rout: Volatility and Bias
Risk management lessons from an AI-fund drawdown: leverage, volatility, short track records, position sizing and decision bias. Includes a precommitted drawdown-review framework.
Risk management lessons from an AI-fund drawdown: leverage, volatility, short track records, position sizing and decision bias. Includes a precommitted drawdown-review framework.
Backtest 332 runs the US-equities book from Backtest 317 and the world-indices book from Backtest 308 on one shared cash account. Over Jan 2016 – May 2026 the blended portfolio compounded at 34.84% CAGR with a 27.13% max drawdown, versus 14.94% and 33.92% for the S&P 500.
Backtest 317 tests a long-only daily-bar systematic strategy on 4,176 US-listed equities from March 2016 to April 2026. It compounds to 30.4% CAGR versus 13.8% for the S&P 500, with a 17.7% max drawdown while holding positions only 78% of the time.
Backtest 308 tests a long-only daily-bar systematic strategy on 1,006 world major-index constituents, now including KOSPI 50, from February 2016 to April 2026. It compounds to 32.7% CAGR versus 14.2% for the S&P 500, with a 14.36% max drawdown.
Comparing fixed fractional, Kelly and sigma-targeting position sizing on a multi-version KreamEdge backtest bench. The bench cannot cleanly isolate a sizing edge.
We built a 3-state SPX percent-rank market-regime gate and tested it on 14 systematic quant backtests. None improved. The base rates and the structural reasons why – plus what to try next.
Long-only daily backtest on 956 world major-index constituents (20 Jan 2016 → 3 Oct 2025): CAGR 27.6% vs SPX 13.1%, Max DD -15.55%, Sharpe 0.89, 654 trades. Backtest, not a live signal.
Backtest of the systematic World 1D strategy over 7,666 US and European equities on a daily timeframe across roughly 10 years. It produced a 40% CAGR with a 29% max drawdown, 1.79 Sortino and 1.15 Sharpe over 374 trades – backtested, not live, figures.
Disclaimer. Content on KreamEdge is published for informational and educational purposes only. Backtest results and strategy outputs do not constitute personalised investment advice nor an invitation to buy or sell any financial instrument. Past performance is not indicative of future results. Financial markets involve risk of partial or total capital loss. Each reader is solely responsible for their investment decisions and risk management. Full bilingual disclaimer on the community page.