newbie question about rebalance and position

[finlabhk]

I am a newbie on doing backtesting. I dont know how to do the backtesting on rebalance my portfolio.

Strategy Description:

select a pool of targets

symbols = ['BTC-USD', 'ETH-USD', 'XRP-USD', 'BNB-USD', 'BCH-USD', 'LTC-USD']
start = '2020-01-01 UTC'  # crypto is UTC
end = '2020-09-01 UTC'
ohlcv = vbt.YFData.download(symbols, start=start, end=end).concat()
close = ohlcv['Close']

find the target which raise over 10% in last 5 days, if there is more than 3 targets, only top 3 targets will be selected. Each target is only allowed to invest 50% of portfolio. Rebalance the portfolio every week

selected = (close.pct_change(5).rank(axis=1)>3) & (close.pct_change(5)>0.1)
selected = selected.resample('1W').first()
selected = selected.div(selected.sum(axis=1), axis=0).clip(None, 0.5)
comb_pf = vbt.Portfolio.from_orders(close, selected.reindex(close.index), price=close, fees=0.001, slippage=0.001, cash_sharing=True)

But I am so confused about the result:
comb_pf.orders.records_readable

Is there any wrong about the coding?
P.S. Thank you @polakowo for contributing such an amazing library to do backtesting.

[polakowo]

Hi @finlabhk, see this answer on how to approach a rebalancing task. There is also an example under https://vectorbt.dev/api/portfolio/base/#vectorbt.portfolio.base.Portfolio.from_orders if you already know the weights. In your example, you need to specify group_by=True or any other group. Your close and any other array must contain all assets as columns.