runMultipleBacktests: Run Multiple Portfolio Backtests and Plot
In PCRA: Companion to Portfolio Construction and Risk Analysis

runMultipleBacktests

R Documentation

Run Multiple Portfolio Backtests and Plot

Description

Runs n_simulations independent portfolio backtests, each time drawing a random subset of stocks from stock_list. For every simulation a full backtest is executed via runPortfolioBacktest(), and an individual plot can optionally be saved. After all simulations finish, the function averages the cumulative return series across simulations and can save a summary plot of those averaged returns.

Usage

runMultipleBacktests(
  n_simulations,
  portfolio_size,
  seed = NULL,
  return_portfolio,
  stock_list,
  buildPortfolios = buildPortfolios,
  market_return = NULL,
  rebalance_on = NULL,
  rolling_window = NULL,
  optimize_method = "CVXR",
  moment_list = NULL,
  save_plot = TRUE,
  plot_path = "./",
  plot_name = "backtest",
  plot_main = NULL,
  plotType = "both",
  save_avg_plot = TRUE,
  avg_plot_path = "./",
  avg_plot_name = "avg_backtest",
  avg_plot_main = NULL,
  avgPlotType = "both",
  colorSet = NULL,
  ltySet = NULL,
  lwdSet = NULL
)

Arguments

`n_simulations`	Integer. Number of independent backtest simulations to run.
`portfolio_size`	Integer. Number of stocks to randomly sample from `stock_list` for each simulation.
`seed`	Integer. Random seed passed to `set.seed()` before the simulation loop for reproducibility.
`return_portfolio`	An `xts` matrix of asset returns. Column names must match the assets in `stock_list`.
`stock_list`	Character vector. Universe of stock tickers from which `portfolio_size` tickers are drawn at random in each simulation.
`buildPortfolios`	A function that accepts a character vector of selected stock tickers and returns a named list of `portfolio.spec` objects, one per strategy. The names of the list elements are used as strategy labels in plot legends and all other outputs. Typically built with `buildPortfolios` and can be customized.
`market_return`	An `xts` single-column object of benchmark returns.
`rebalance_on`	Character string passed to `optimize.portfolio.rebalancing()`. See `endpoints` for valid names.
`rolling_window`	Positive integer. Length of the rolling estimation window in periods.
`optimize_method`	Character string specifying the solver. Default `"CVXR"`.
`moment_list`	If different moment functions are passed into multiple GMV portfolios, please define each moment function via this parameter. For the portfolio that do not require moment function, please pass NULL. Example: `list('custom.covRob.Rocke', NULL, NULL)`.
`save_plot`	Logical. Whether to save the plot to a PNG file. Default `TRUE`.
`plot_path`	Character string. Full file path for the each simulation output. Required when `save_plot = TRUE`.
`plot_name`	Plot name for each simulation output. Default `"backtest"`
`plot_main`	Plot title for each simulation PNG output.
`plotType`	"cumRet", "drawdown", or the default is "both"
`save_avg_plot`	Logical. Whether to save the the average cumulative returns plot to a PNG file. Default `TRUE`.
`avg_plot_path`	Character string. Full file path for the average simulation output. Required when `save_avg_plot = TRUE`.
`avg_plot_name`	Plot name for average simulation output. Default `"avg_backtest"`.
`avg_plot_main`	Plot title for the average simulation PNG output.
`avgPlotType`	"cumRet", "drawdown", or the default is "both"
`colorSet`	Optional character vector of colors passed to `backtest.plot()`.
`ltySet`	Optional integer vector of line types passed to `backtest.plot()`.
`lwdSet`	Optional integer vector of line width passed to `backtest.plot()`.

Value

A named list with four elements:

results: A list of length n_simulations. Each element contains the returns component returned by runPortfolioBacktest() for that simulation.
selected_stocks_all: A list of length n_simulations. Each element is a character vector of the tickers chosen for that simulation.
avg_algorithm_returns: An xts object containing the element-wise average of all simulations' algorithm return series.
avg_cumulative_return: An xts object containing the element-wise average of all simulations' cumulative return series.