Moonshot Backtest¶
Notebooks provide a great way to develop your strategy logic interactively, but Moonshot backtests and parameter scans are often superior tools for running the code and analyzing performance, especially for large datasets that may not fit into memory.
Because Moonshot is pandas-based, your interactive strategy code can readily be transferred to a Moonshot strategy. The file qval.py contains our transferred strategy code. To facilitate parameter scans, we've factored out some hard coded parameters and stored them as class attributes on the strategy (see for example VALUE_TOP_N_PCT and QUALITY_TOP_N_PCT).
Install strategy file¶
Moonshot looks for strategies inside the "moonshot" directory, so execute the following cell to "install" the strategy by moving the file to that location:
The
!sytax below lets us execute terminal commands from inside the notebook.
# make directory if doesn't exist
!mkdir -p /codeload/moonshot
!mv qval.py /codeload/moonshot/
Moonshot reloads this file each time you run a backtest. Thus, any edits you make to qval.py will automatically be reflected in the subsequent backtest.
Run backtest¶
Note the CODE we assigned to our strategy, "qval":
class QuantitativeValue(Moonshot):
CODE = "qval"
This is how we refer to this strategy for running a backtest.
Since this strategy requires a large amount of fundamental and price data, we probably can't load a 10-20 year backtest into memory. We use the segment parameter to tell Moonshot to run the backtest in 1-year segments and concatenate the results. ("A" is the pandas offset alias for annual frequency.)
We've added Jupyter's
%timemagic to provide a gauge of total runtime for this backtest.TIP: open the detailed logs while running the backtest to monitor moonshot's progress through the backtest segments.
from quantrocket.moonshot import backtest
%time backtest("qval", start_date="2010-01-01", end_date="2018-01-01", segment="A", filepath_or_buffer="qval_backtest.csv")
from moonchart import Tearsheet
Tearsheet.from_moonshot_csv("qval_backtest.csv")
Pyfolio¶
import pyfolio as pf
pf.from_moonshot_csv("qval_backtest.csv")
