QuantStats

By calling qs.extend_pandas(), you can attach Monte Carlo functionality directly to pandas Series objects.

qs.extend_pandas()

# run simulation directly on a returns Series
mc = returns.montecarlo(sims=1000, bust=-0.15, goal=0.30)

# plot directly
returns.plot_montecarlo(sims=500)

To use QuantStats, import the library and call qs.extend_pandas() to add performance metrics directly to pandas Series objects. You can then download returns using qs.utils.download_returns() and calculate metrics like the Sharpe ratio.

%matplotlib inline
import quantstats as qs

# extend pandas functionality with metrics, etc.
qs.extend_pandas()

# fetch the daily returns for a stock
stock = qs.utils.download_returns('META')

# show sharpe ratio using the stats module
qs.stats.sharpe(stock)

# or using the extended pandas method
stock.sharpe()

You can install QuantStats using either pip or conda.

$ pip install quantstats --upgrade --no-cache-dir

# Or using conda
$ conda install -c ranaroussi quantstats

QuantStats includes built-in Monte Carlo simulations for probabilistic risk analysis. Use qs.stats.montecarlo() to run simulations and analyze bust and goal probabilities.

import quantstats as qs

# returns is a pandas Series of daily returns
mc = qs.stats.montecarlo(returns, sims=1000, bust=-0.20, goal=0.50)
print(f"Bust probability: {mc.bust_probability:.1%}")
print(f"Goal probability: {mc.goal_probability:.1%}")
mc.plot()

You can visualize simulation results using qs.plots or directly from a MonteCarloResult object.

• Plot all paths: Use qs.plots.montecarlo(returns, sims=500) or mc.plot().
• Plot terminal distribution: Use qs.plots.montecarlo_distribution(returns, sims=500) to see the distribution of final values.

# Plot all simulation paths
qs.plots.montecarlo(returns, sims=500, seed=42)
# or from an existing result
mc.plot()

# Plot terminal value distribution
qs.plots.montecarlo_distribution(returns, sims=500, seed=42)

Use qs.plots.snapshot() to create a comprehensive visual summary of stock performance. This can also be called directly on a pandas Series if qs.extend_pandas() has been used.

import quantstats as qs

# Using the plots module
qs.plots.snapshot(stock, title='Facebook Performance', show=True)

# Or using the extended pandas method
stock.plot_snapshot(title='Facebook Performance', show=True)

The MonteCarloResult object provides tools to analyze risk across paths:

• Max Drawdown Distribution: Access mc.maxdd for a dictionary of statistics (min, max, mean, median, std, percentiles) regarding the maximum drawdown of each path.
• Confidence Bands: Use mc.confidence_band(level) to get the lower and upper bounds for a given confidence level (e.g., 0.95).
• Percentile Paths: Use mc.percentile(p) to retrieve a specific percentile path (e.g., mc.percentile(10) for the 10th percentile).

# Max drawdown distribution
print(mc.maxdd)

# 95% confidence band
lower, upper = mc.confidence_band(0.95)

# specific percentile path
p10 = mc.percentile(10)

Perform probabilistic risk analysis by shuffling historical returns to simulate thousands of possible outcomes. Use qs.stats.montecarlo to generate a MonteCarloResult object containing terminal statistics, drawdown distributions, and path data.

import quantstats as qs

# fetch daily returns
returns = qs.utils.download_returns('SPY')

# run Monte Carlo simulation with 1000 paths
mc = qs.stats.montecarlo(returns, sims=1000, seed=42)

# view terminal value statistics
print(mc.stats)

When running a simulation, you can specify bust (a drawdown threshold) and goal (a return threshold) to calculate the probability of hitting these specific levels during the simulated paths.

# Set bust threshold at -20% drawdown, goal at +50% return
mc = qs.stats.montecarlo(returns, sims=1000, bust=-0.20, goal=0.50, seed=42)

print(f"Probability of bust: {mc.bust_probability:.1%}")
print(f"Probability of reaching goal: {mc.goal_probability:.1%}")

QuantStats provides helper functions to get the distribution of key performance metrics across all simulated paths.

# Sharpe ratio distribution
sharpe_dist = qs.stats.montecarlo_sharpe(returns, sims=1000)

# Max drawdown distribution
dd_dist = qs.stats.montecarlo_drawdown(returns, sims=1000)

# CAGR distribution
cagr_dist = qs.stats.montecarlo_cagr(returns, sims=1000)