How it works

Monte Carlo simulation is built from standard, transparent statistics on your own basket. It measures how each stock has behaved, adjusts those measurements for the scenario you pick, rolls thousands of randomized future paths, and then summarizes where the paths landed. This page explains the method at a conceptual level, with the standard finance math shown in full. The tool is available to Free, Pro Lite, and Pro accounts.

What it shows

The process has four steps: measure each stock's behavior, apply the chosen scenario, roll thousands of future paths, and summarize the outcomes. The sections below walk through each step.

Measure each stock's behavior

For every holding, the tool reads the daily price history over your lookback window and computes the average daily return, the volatility (how much daily returns scatter), and the correlation between holdings (how strongly they move together). The daily return uses the log change in the split-and-dividend-adjusted closing price.

r = ln(P today / P yesterday)

In this formula, r is the daily return and P is the adjusted closing price, measured each trading day. For example, a stock at 100.00 yesterday and 101.00 today has r equal to ln(101 divided by 100), which is 0.00995, about a 1.0 percent day. A stock with only a short price history gets a less reliable estimate, which the Holdings table flags.

Apply the chosen scenario

Before simulating, the selected Risk Scenario scales the estimated return and the volatility up or down to model a calmer or harsher regime. Current Conditions leaves them as measured. The other scenarios raise or lower returns and volatility. The 2008 Crisis preset both lowers returns and sharply raises volatility, and it uses a heavier-tailed draw to mimic shocks. These scenarios are uniform stress settings, not forecasts of any specific event.

Roll thousands of future paths

For each path, the tool draws correlated random daily returns for every holding, matching the measured correlation, blends them by your weights, and compounds the portfolio forward one day at a time across the horizon.

value next = value now × exp(portfolio daily return)

Here value now and value next are the portfolio's dollar value on consecutive trading days, and portfolio daily return is the weighted blend of that day's drawn holding returns. With contributions or withdrawals scheduled, cash is added or removed on the chosen cadence. With rebalancing on, weights are restored to target on the chosen cadence, and any commission, slippage, and dividend tax are charged on those trades. The number of paths equals your Simulation Quality, from 5,000 to 100,000, and each year is treated as 252 trading days.

Summarize the outcomes

After every path finishes, the tool sorts the ending values to read off percentiles. The median is the 50th percentile, and the bands come from the 5th, 25th, 75th, and 95th. It counts how often each result happened to build the histogram, and it computes the risk figures.

Probability of Loss is the count of paths ending below your starting capital, divided by the total number of paths. Win Rate is one minus that.
Target Hit Probability is the count of paths ending at or above your target, divided by the total.
Value at Risk at 95 percent is the 5th-percentile ending value minus your starting capital. Conditional VaR averages the outcomes in that worst 5 percent.
The risk-adjusted ratios divide return by a measure of risk, using an assumed risk-free rate shown in the Run Summary.

For a worked example, if 18,000 of 100,000 paths finish below your start, the Probability of Loss is 18 percent and the Win Rate is 82 percent. If the 5th-percentile ending value is 82,000 on a 100,000 start, the Value at Risk at 95 percent is minus 18,000, meaning a 5 percent chance of losing more than 18,000. These are probabilities inside the model's assumptions, not guarantees about markets.

Data and timing

The simulation runs on daily split-and-dividend-adjusted closing prices. Market data comes from AlphaVantage, with Nasdaq-listed quotes delivered through AlphaVantage, and correlation is measured against a broad US market index. Prices refresh each trading day after the close, so each run uses history through the most recent completed trading day. A run finishes in roughly 8 seconds to 3 minutes, depending on the chosen quality.

Use cases

Choose a longer lookback to average over more market conditions, or a shorter one to weight recent behavior.
Switch scenarios to stress the same basket under a calmer or harsher regime.
Raise the simulation quality to steady the percentile estimates before a decision.

Limitations and disclosures

The simulation assumes the future behaves statistically like the chosen historical window, so it cannot anticipate regime changes, liquidity gaps, or events outside that window. Returns are modeled as a smooth random process with stable statistics, so sudden discontinuities are not captured, and the risk-free rate in the ratios is a fixed assumption. Outcomes are modeled probabilities, not guarantees, and real markets can move outside the simulated range.

GNG Research provides equity research and educational tools, not investment advice. Nothing on the platform is a recommendation to buy or sell any security. Do your own research and consider your circumstances before making any investment decision.