This is the 2D probability-distribution counterpart to the 3D capital-stack simulator. Instead of stacking 3D boxes, it draws the full binomial probability mass function of the portfolio's default count directly — a statistical/parameter-space view of the same waterfall math, not a flattened camera angle on the same scene.
The portfolio holds N = 30 projects of equal size. Each defaults independently with probability p, so the number of defaults k follows a binomial distribution:
P(k defaults) = C(N,k) · p^k · (1-p)^(N-k)
Portfolio loss L(k) = k · (100/N) · LGD (% of committed capital)
Losses cascade up the stack bottom-up:
First-loss absorbs min(L, Lf)
Mezzanine absorbs min(max(L − Lf, 0), Lm)
Senior(private) hit min(max(L − Lf − Lm, 0), 100 − Lf − Lm)
The bar chart bins every possible k from 0 to 30 by its exact probability, and colors each bar by which tranche that outcome would breach. The horizontal gauge below is a flat 2D capital-stack bar (not a 3D box): tranche widths are proportional to their share of capital, and the red fill shows how far the currently-simulated year's loss climbs across it. The row of 30 squares shows exactly which individual projects defaulted this year. Clicking Simulate Year repeatedly builds an empirical (Monte Carlo) frequency histogram, drawn as a thin outline over the theoretical bars, so you can watch simulation converge to the closed-form probability curve.
The catalytic ratio = Senior size ÷ First-loss size measures how many dollars of private capital one dollar of public first-loss capital mobilizes.
- First-loss / mezzanine sliders — resize the junior tranches; the remainder is automatically the private senior tranche.
- Default probability / LGD sliders — set the underlying project risk that feeds the binomial model.
- Simulate Year — draws one random realization of defaults, updates the project strip and stack gauge, and adds one sample to the running Monte Carlo tally.