Why Your Backtest Is Lying to You: Data Biases That Ruin Equity Research
Survivorship bias, look-ahead bias, point-in-time data, and the rest of the subtle ways your beautiful backtest is overstating returns. With concrete fixes for each.
Investing tools, market essays, and developer-to-investor crossovers — tested or argued, never speculated.
8 articles
Survivorship bias, look-ahead bias, point-in-time data, and the rest of the subtle ways your beautiful backtest is overstating returns. With concrete fixes for each.
Compound interest sounds like magic. The math is mundane and the most common rules of thumb are subtly wrong. Here's a calmer take on what compounding actually delivers.
The three labels overlap more than the marketing suggests. A developer-friendly breakdown of structure, tax treatment, and the cases where the difference actually changes your decision.
Polygon and Alpha Vantage are the two common entry points for developer-side stock data. A side-by-side on free tiers, coverage, and developer experience — without the trading-edge hype.
Most stock-data tutorials skip past the one free, public, no-rate-limit source for US company filings. Here's how to read fundamentals straight from SEC EDGAR, JSON-first.
Diversification reduces idiosyncratic risk, not market risk. When the distinction matters, what correlation looks like under stress, and why 60/40 isn't a free lunch.
Rules-based investing isn't AI, isn't HFT, and isn't a guarantee. It's a behavioral discipline: decide the action before the event. Here's what counts as a rule, what doesn't, and how it actually fails.
Tools, evidence, and disclosure are the same job whether the topic is a code editor or a brokerage account. Here's the editorial line for the new finance section.
One email every Friday with the tools worth your time. No spam, unsubscribe anytime.