This. Even if your annual returns are normally distributed, by multiplying the returns together, the ending balance is certainly not normally distributed, so we don't expect the median to be the same as your deterministic model.I think your P50 would only match your deterministic value if your uncertainties are normal distributions. I think most uncertainties are closer to beta distributions that skew in one direction.
Statistics: Posted by AnEngineer — Tue Dec 23, 2025 6:30 am — Replies 16 — Views 726