Backtest Period
Markets Traded
Equities
Maximum Drawdown
Period of Rebalancing
Yearly
Return (Annual)
Sharpe Ratio
Standard Deviation (Annual)
Original paper
SSRN-id2494412.pdf970.7KB
Trading rules
- Investment universe: 500 largest non-financial stocks with available gross profits-to-assets and book-to-market ratios
- Annual ranking: Rank stocks based on gross profits-to-assets and book-to-market ratios (1=lowest, 500=highest)
- Portfolio construction: At the end of each June, buy $1 of each of the top 150 stocks with highest combined profitability and value ranks; short $1 of each of the bottom 150 stocks with lowest combined ranks
- Rebalancing: Rebalance the portfolio yearly
Python code
Backtrader
import backtrader as bt
class ProfitValueStrategy(bt.Strategy):
def __init__(self):
self.rank_size = 500
self.top_bottom_size = 150
def next(self):
if self.datetime.date().month == 6 and self.datetime.date().day == 30:
self.rank_stocks()
self.construct_portfolio()
self.rebalance()
def rank_stocks(self):
stocks = self.datas
ranks = []
for stock in stocks:
gross_profit_to_assets = stock.gross_profit / stock.total_assets
book_to_market = stock.book_value / stock.market_cap
ranks.append((stock, gross_profit_to_assets, book_to_market))
self.ranks = sorted(ranks, key=lambda x: (x[1], x[2]), reverse=True)[:self.rank_size]
def construct_portfolio(self):
self.long_stocks = [stock[0] for stock in self.ranks[:self.top_bottom_size]]
self.short_stocks = [stock[0] for stock in self.ranks[-self.top_bottom_size:]]
def rebalance(self):
self.broker.set_cash(0)
for stock in self.long_stocks:
self.order_target_value(stock, target=1)
for stock in self.short_stocks:
self.order_target_value(stock, target=-1)
cerebro = bt.Cerebro()
# Add data to cerebro (not shown)
cerebro.addstrategy(ProfitValueStrategy)
cerebro.broker.set_coc(True)
results = cerebro.run()This code is a sample implementation of the trading rules using the Backtrader library. Please note that you would need to add your stock data to the ‘cerebro’ object and ensure that the stock data includes the required attributes (gross_profit, total_assets, book_value, and market_cap) for the ranking calculations.