Limitations & caveats

The simulation is a powerful lens on sequence risk. It is not a crystal ball, and it has specific, knowable blind spots. Knowing them is part of using it.

It can only rearrange the past

Paths are resampled from your portfolio's daily P/L over the selected backtest window. The simulation can reorder those days in thousands of ways — it cannot invent days the window never contained. If the window holds no crash, no simulated future crashes (that's exactly what the stress injector is for — use it). The flip side: a window dominated by one regime produces futures dominated by that regime.

Days are drawn independently

Resampling draws days with replacement, independently — but real markets have memory: volatile days cluster, losing streaks in 0DTE strategies often come from related market conditions, not independent bad luck. Independent resampling therefore tends to understate clustering — real drawdowns can run deeper than the simulated distribution suggests, because real bad days arrive in convoys.

The portfolio is frozen

Every path assumes you trade the same strategies at the same contract counts for the whole horizon — no scaling down after a drawdown, no stopping, no rebalancing. Real behavior (yours) will differ, in both directions.

Probability of Profit is not a forecast

It's the share of resampled histories ending above starting capital — a statement about the backtest window's composition, not about next year. The app itself caps the display and says as much.

How to use it anyway

• Treat the drawdown distribution as the primary output; the return cards are dessert.
• Run it with stress on (1–2 events/yr, Black Swan mode for defined-risk worst case) and size so that that version is survivable.
• Re-run after every meaningful portfolio change — the simulation is cheap; finding out live is not.

All results are hypothetical, for research and education. Nothing here is a recommendation or a projection of future results.