Original data-driven studies in personal finance math. Every study includes its full methodology, PRNG seed, raw scenario distribution, and limitations — reproduce any number we publish, or fork the dataset for your own analysis.
Deterministic Monte Carlo of 10,000 30-year portfolios. Median, percentile spread, contribution vs starting-balance attribution, and the year interest beats contributions.
Read study →1,000 multi-debt profiles simulated under both orderings. Median interest gap, payoff-time gap, and the profile types where each method actually wins.
Read study →Swept 5,000 (mortgage rate × equity return × horizon) cells to find the exact crossover where prepay beats investing the same $500/month.
Read study →2,500 borrower profiles (balance × salary × profession) under standard 10-yr, IDR + PSLF/RAP forgiveness, and private refi. Profession beats balance — and the RAP tax bomb is bigger than people think.
Read study →Modeled 10,000 finance-publisher traffic mixes against the May 2026 AI Overviews rollout. Tool-required pages lose ~12% of traffic; pure definition content loses ~64%. The protection is the tool, not the content.
Read study →Original composite 0-100 index per state combining real income surplus (income ÷ COL), state tax burden, and Shiller CAPE market entry climate. Tennessee leads at 66.7; Hawaii trails at 43.3. Free JSON dataset.
Read study →10,000 investor profiles simulated under on-time vs delayed starts. Quantifies the median dollar cost of waiting 1, 5, and 10 years. Cost compounds non-linearly with delay.
Read study →10,000 investor profiles compared across five fee tiers (0.03% to 1.50%). Quantifies what each basis point in expense ratio actually costs by retirement — including the long-horizon penalty.
Read study →