Monte Carlo Simulation for Traders: What 1,000 Runs Tell You About Your Edge

Your backtest lied to you (kind of)

Here's a thing that took me an embarrassingly long time to figure out: a backtest only shows you one path through historical data. One specific sequence of wins and losses, in one specific order, during one specific stretch of market conditions. And that's the path you're using to bet real money.

But you're not going to trade that exact sequence again. That's not how markets work. What if your three worst trades happened back-to-back instead of being spread across three different months? What if the hot streak you had in March moved to December, and meanwhile you spent the first quarter grinding through your drawdown?

A backtest answers: "What DID happen?" Monte Carlo simulation answers a much more useful question: "Given my stats, what COULD happen?"

That second question is the one that keeps your funded accounts alive.

What Monte Carlo actually does (in plain English)

The name sounds intimidating. It's not. Here's what it actually does, step by step.

It takes your trading stats -- win rate, average win, average loss -- and runs 1,000 random simulations of 1,000 trades each. Every simulation uses your exact same numbers. Same 55% win rate. Same $200 average win and $150 average loss. Nothing changes except the order.

Each simulation is like a parallel universe where you trade the same strategy, with the same edge, but the sequence of wins and losses gets shuffled randomly. In universe #1, you start with five winners in a row. In universe #472, you start with eight losers. In universe #999, it's a choppy mix. All of them have the same overall win rate -- but the paths they take to get there are wildly different.

When you plot all 1,000 equity curves on top of each other, you get what traders call the "spaghetti chart." And it tells you something a single backtest never could: the range of outcomes your edge can produce.

Reading the spaghetti chart

The first time you see a spaghetti chart, it looks like chaos. A thousand squiggly lines going in roughly the same direction but spreading out like a fan. Here's how to actually read it.

• The spread -- how wide the fan gets -- shows you how much randomness can affect your outcome, even with a positive edge. Wider spread means more variance. More variance means more opportunity for a bad streak to wreck you before the math works out.
• Paths that go below zero -- those are simulated blowups. Your strategy has a positive expectancy, but in those particular sequences, you hit drawdown before the wins caught up. If 8% of paths go below zero, that's your blow-up probability.
• The blue median line -- the thick line running through the middle -- represents the 50th percentile. Half of all simulated outcomes finished above it, half below. This is your most realistic "expected" result, way more reliable than any single backtest.
• Dense clusters -- where the lines bunch together -- those are your likely outcome zones. If most paths cluster between +$20K and +$40K, that's your realistic range. The outliers above and below are possible but less probable.

The stats that matter

The spaghetti chart is the visual. But the real value comes from the numbers the simulation spits out. Here are the ones I actually look at every week:

• Median final P&L -- the 50th percentile outcome. Not the average (which gets skewed by outliers), but the middle. This is what you should realistically expect if you execute your strategy consistently over the sample size.
• Max drawdown (median) -- how deep the typical hole gets before you recover. If your median max drawdown is $2,800 and your account has $3,000 in available drawdown, you have almost no margin for error. That's a problem.
• Worst-case scenario -- the bottom 5% of runs. This is the "things went really badly" number. Not the absolute worst (that's often an outlier), but the realistic worst case. If the 5th percentile outcome is -$4,000, you need to know that.
• Blow-up probability -- what percentage of simulated runs went below your drawdown limit. Under 5% is my personal comfort zone. Above 10% and I'm reducing position size immediately.
• Win/loss streaks -- the longest losing streak across all 1,000 simulations. This one's for your head. If the sim says you could realistically hit 12 losers in a row, and you've never emotionally prepared for that, you're going to panic at loss #6 and blow up your process.

What the simulation doesn't tell you

I want to be honest about this because I see too many traders treat Monte Carlo like a crystal ball. It's not. It's a stress test. Here's what it can't do:

How I use it every week

Here's my actual workflow. Nothing fancy. Takes about five minutes.

Every Sunday night, I update my trading stats from the previous week. Win rate, average winner, average loser. If I've made changes to my strategy -- new entry criteria, different stop placement, anything -- I rerun the Monte Carlo simulation with the updated numbers.

I'm looking at three things:

1. Blow-up rate under 5%. If it creeps above that, I'm either oversized or my edge has degraded. Either way, I'm scaling down before Monday morning.
2. Worst-case max drawdown vs. my real drawdown. I take the 5th percentile max drawdown number from the sim and compare it to my actual available drawdown across my accounts. If the sim says worst case is $2,500 and I've only got $1,800 of drawdown left on an account, that account sits out until the buffer is healthier.
3. The 60% rule. If the worst-case max drawdown exceeds 60% of my available drawdown on any account, I reduce risk on that account. No exceptions. This is the rule that stopped me from blowing accounts after I built this tool.

The whole process takes less time than making coffee. And it has saved me more money than any indicator, course, or trading room ever did.

The quarter-Kelly trick

Quick sidebar on something the FundedSizer Monte Carlo sim does automatically that's worth understanding.

The Kelly Criterion is a formula that tells you the mathematically optimal bet size given your win rate and reward-to-risk ratio. The formula is simple: Kelly % = win rate - (1 - win rate) / R, where R is your average win divided by your average loss.

Problem is, full Kelly is insane. It's the theoretically optimal bet size for maximizing long-term geometric growth, but in practice it produces stomach-churning drawdowns. We're talking 40-60% of your bankroll swings. Nobody trades like that, and prop firm drawdown rules wouldn't let you even if you wanted to.

That's why most professional traders and quants use quarter Kelly -- you take the Kelly percentage and divide by four. It captures most of the growth (roughly 75% of full Kelly's return) while cutting drawdown by about half. It's the sweet spot between growth and survival.

When you run the Monte Carlo simulation in FundedSizer, it auto-calculates your quarter-Kelly risk based on the stats you input. You don't have to think about it. But now you know what's happening under the hood and why the suggested risk size might be more conservative than you expected. That conservatism is the point.

For a deeper understanding of how win rate, R-multiple, and expectancy interact, check out the expectancy calculator. It makes the relationship between these numbers a lot more intuitive.

So should you bother?

A backtest shows you one path. The one that already happened. Monte Carlo shows you a thousand paths that could happen, and whether your drawdown survives the ugly ones.

If you're trading prop firm accounts, this isn't extra credit. Your drawdown is a hard line. Cross it and you're resetting another $150 eval. I've done that enough times to know.

Run the sim before Monday. It takes five minutes and it'll tell you more about your position sizing than your last month of backtesting did.