The sample dataset has been replaced with real-world data. Pricing data has been sourced from the Tiingo Stock API, and financial data has been sourced from EDGAR. The dataset includes size and value factors for most of the stocks in the S&P 500 for the period June-August 2020.
The sample shiny app has multiple improvements, including the ability to load simulation results and improved visualizations.
The PortOpt class has been exported. This class can be used to solve single portfolio optimization problems.
Writing html reports for a simulation based on an Rmd template is now possible using the writeReport method of the Simulation class.
Delistings can now be entered with a delisting return.
Average volume has been changed to a market value measure from a shares measure. This makes it much easier to provide a measure of ADV (average daily volume) in a reference currency for setting trading and position limit constraints. By default the column is called rc_vol.
Added input_dates parameter to Simulation class constructor. The date vector passed to input_dates contols the dates on which input data (e.g., alpha and risk factors) is updated.
Added force-trim feature to trim back positions that grow too large. Setting the simulator configuration parameter force_trim_factor to a numeric value X will cause orders to be created that trim positions back to X times their maximum allowable size.
Added force-exit feature for non-investable securities. Setting the simulator configuration parameter force_exit_non_investable to TRUE will cause exit orders to be created for positions not in the universe during a simulation.
Added feature to limit LMV/SMV increase/decrease on a single day. This feature is useful when ramping up the portfolio to its target LMV/SMV at the beginning of the simulation. For example, setting the simulator configuration parameter max_weight_change to 0.1 will limit LMV/SMV increase to 10% of the target LMV/SMV per day.
Now saving gross, long, and short exposures in addition to net.
Standardized simulation result data and plotting methods to have a strategy_name parameter where appropriate.
Improved handling of delistings.
Added new plotting methods for the Simulation class: plotContribution, plotTurnover, plotUniverseSize, and plotNonInvestablePct.
Added overallReturnsByMonthDf method for the Simulation class.
Fixed recording of loosened constraints during simulation.
Fixed loading of configuration information in readFeather.
The example in the package vignette now uses the sample datasets included in the package.
overallStatsDf() includes a max drawdown entry.
Rsymphony dependency has been moved from Imports to Suggests to make installation easier.
Example shiny application now uses the sample datasets provided with the package, and can now be run directly by calling the new function example_shiny_app().
New function example_strategy_config() provides an easy way to grab the sample configuration used in examples and the shiny app.