I used to trust a smooth backtest line the way you'd trust a memory — like it proved something. It doesn't. Quick note first: this is educational commentary, not personalized financial advice.

Here's the leak I finally fixed. A backtest doesn't answer does my edge make money? It answers a much smaller question: did this exact order of trades, on this exact history, make money? Shuffle the order of the same wins and losses and the ending balance can match — but the drawdowns, the moments you'd have quit, are completely different.

Think of a coin that pays $1.10 on heads and loses $1 on tails. Real edge. Flip it 100 times and plot the balance. Do it again — totally different line. Your backtest is just one of those lines. One life the market happened to live.

So take your actual trade list and shuffle it ten thousand times, replaying the equity curve each time. Same edge, same trades — only the sequence changes. Now you have a distribution, not a destiny. Sort the endings: the median is a realistic expectation, the fifth percentile is the bad-but-plausible year. And the max drawdown across all those paths? A 12% worst case in your single run can become 30% or 40% once losses cluster in a crueler order. That deeper number is the one you actually have to survive.

The line I keep coming back to: your edge decides whether you make money over infinite trades; your position size decides whether you're still trading when infinity arrives.

What a disciplined trader does: never size off a single equity curve. Export the trade list, reshuffle it thousands of times, and set position size so that even the fifth-percentile drawdown is one you can stomach without flinching. Ask can I survive the unlucky order — not how good was the lucky one.

One caveat: this assumes trades are independent and your edge is stable. Markets cluster and regimes shift, so treat Monte Carlo as a stress test, not a crystal ball.

Full walkthrough here: https://youtu.be/M18sv0jf39c

— Paragon