What is Monte Carlo simulation? Understanding the math behind outcomes

The core idea Monte Carlo simulation is a computational technique that runs thousands of random scenarios to explore the range of possible outcomes. Rather than assuming the market delivers one expected return every year, it models realistic volatility: some years spike up, some crash down, and the order is random.

By the end, you see a *distribution* of outcomes instead of a single forecast. Some simulations hit your goal; others fall short. The percentiles show you the landscape: downside risk, upside potential, and most likely middle outcome.

How it actually works Here's a simplified version of what happens under the hood:

1. You set the inputs: initial investment, monthly contribution, expected annual return (e.g., 12%), volatility or standard deviation (e.g., 18%), and time horizon (e.g., 20 years).

2. For each scenario: the simulator runs 240 months (20 years × 12). Each month: - Your balance grows by a randomized return drawn from a normal distribution (a bell curve). - The return centers on your expected return but has a realistic spread (volatility). - Some months might see +5%, others -3%; over thousands of scenarios, the mix of good and bad months is naturally modeled.

3. After all months: the final portfolio value for that scenario is recorded.

4. Repeat thousands of times: 10,000 scenarios means 10,000 different paths to a final value, each realistic given your assumptions.

5. Sort and extract percentiles: sort all 10,000 final values from smallest to largest. The 10th percentile is the value at position 1,000 (10% of 10,000). The 50th is the median. The 90th is near the top.

Why normal distribution? Historical market returns roughly follow a bell curve: most outcomes cluster around the average, extreme gains and losses are rarer but real. The normal distribution captures this. Each month's random return is drawn from this curve, scaled by volatility (wider curve = more unpredictable markets; tighter curve = steadier markets).

Volatility: the key variable Volatility (standard deviation) determines how spread out the outcomes are. Historical equity returns in India average around 15-20% volatility: - A market with 10% volatility is stable; outcomes cluster tightly around the expected return. - A market with 20% volatility is wild; outcomes spread widely. Some years +40%, others -20%, even if the long-run average is 12%.

Getting volatility right is critical. Underestimate it and your worst-case scenario is far worse than your plan expects. Overestimate it and your plan looks doomed when it is actually safe.

What Monte Carlo assumes (and doesn't capture) Monte Carlo assumes returns follow a normal distribution, repeating the past. But real markets can behave very differently: - Black swan events: rare crashes that fall outside historical norms. A 50% market crash in one month is extremely unlikely by normal distribution math, but it has happened. - Regime shifts: the return and volatility of the 1990s might not repeat in the 2020s. Monte Carlo can't predict structural change. - Rebalancing: the simulation can model whether you rebalance or not, but real-world rebalancing is lumpy and taxes have an impact.

Despite these limits, Monte Carlo is far more realistic than assuming steady returns every year. It captures sequence risk and volatility — the two biggest threats to a long-term plan.

Percentiles: what they mean - 10th percentile: 90% of simulations end up above this value. This is downside risk; the outcome if you have remarkably bad luck. - 25th percentile: 75% of outcomes beat this. A conservative but realistic target. - 50th percentile (median): the true middle. Half the simulations do better; half worse. - 75th percentile: a strong outcome. Still realistic but on the optimistic side. - 90th percentile: only 10% of simulations beat this. Tempting to plan for, but dangerous.

Most financial advisors suggest targeting a "success rate" of 85-95%, meaning your goal is hit in 85-95% of simulations. This roughly maps to the 25th-50th percentile range depending on your risk tolerance.

Why this beats a simple average A single average (e.g., "you'll have INR 1 crore") hides the full story. Monte Carlo reveals it: in some scenarios you have INR 60 lakh; in others INR 2.5 crore. Planning for the average and then panicking when you hit the 10th percentile is a recipe for selling at exactly the wrong time. Knowing the range in advance lets you stress-test your emotional tolerance and adjust your plan before crisis hits.