Graham Stephan

Macro calls like 'the middle class is about to get screwed' or 'we're going to be broke' are directionally plausible but lack specificity — they cannot be cleanly verified or falsified, inflating apparent accuracy

Only 20 of 38 predictions have been scored; 18 remain pending. The scored subset skews toward older, more resolvable calls, potentially overstating accuracy until pending predictions resolve

With zero market-priced predictions and no alpha computation, the 77.5% accuracy figure cannot be compared to simply buying and holding the S&P 500 — it measures narrative correctness, not investment performance

Several predictions are near-duplicates of the same underlying claim (e.g., bond market crack appears twice, $50T debt projection appears twice, CBO interest projection appears twice), which may artificially inflate scored prediction counts and accuracy

The 2026 macro backdrop — elevated deficits, bond yield volatility, tariff uncertainty — made bearish fiscal calls broadly correct regardless of analytical skill, reducing the signal value of these predictions