I'm questioning how you can incorporate data which does not exist, i.e., volume data on something which does not trade and therefore generates no volume. If it is a key component of your backtest and it is faulty, then the validity of your test is suspect. Don't you agree?