Backtest

Backtest framework for trading algorithms.

Usage

1. Create your strategies.

   • Create strategy_name.jl files that contain the algorithms to be executed at each step.

   • Strategies must take a ::backtest_structs.backtest_instance and a ::IOStream, plus any optional arguments that may be relevant. See strategies/faber_momentum.jl for an example. The strategy must return a ::Float every step, indicating the accrued P/L during the current step.

   • Strategies must interface with the framework through open_trade and close_trade functions.

     • close_trade and open_trade do not update any of the backtest information, instead they return relevant values that must be handled by your strategy. These values are:
       • open_trade -> Tuple{position::position_info, available_cash::Float64, equity::Float64.
       • close_trade -> Tuple{profit_loss::Float64, available_cash::Float64, equity::Float64}.

2. Structure your backtest.

   • Create a ::strategy struct, which contains the starting cash (::Float64), strategy (::Function), and optional arguments (::NamedTuple).

   • The entry point will be run_backtest(strategy::backtest_structs.strategy, file_out::String, in_dir::String, etfs::Vector{String}. It will run the strategy through every member of etfs which has data files inside in_dir.

   • You can also save SVGs that plot equity over the the backtest with plot_data(data::Vector{Float64}, filename::String; args)

   • See examples/faber_bt.jl for more information.