Comparing pervious days close to current last price?

Here is the following backtest, where I have tried to obtain the values and do a basic mathematical operation on them. Unfortunately it looks as though this is not working, as the backtest does not return any results at all.

Could someone point out where I am going wrong please?

class TradeStrategyTest(QCAlgorithm):
   
   def Initialize(self):
       self.SetStartDate(2021,1, 1)  #Set Start Date
       self.SetEndDate(2021,6,1)    #Set End Date
       self.SetCash(10000)           #Set Strategy Cash
       self.AddEquity("SPY", Resolution.Minute)
       self.SetWarmUp(14)
      
       
       self.history_day_close = self.History(["SPY"], 1, Resolution.Daily)
       self.previous_day_close = self.history_day_close.loc["SPY"]['close'].item()
       
       self.history_bar_close = self.History(["SPY"], 1, Resolution.Minute)
       self.previous_bar_close = self.history_bar_close.loc["SPY"]['close'].item()
       self.change_from_close = (self.previous_day_close - self.previous_bar_close)
   def OnData(self, data):
       
       if self.IsWarmingUp:
           return
   
   
       if not self.Portfolio["SPY"].Invested:
           
           if ((self.change_from_close <=-4) and (self.change_from_close >=-10)):
               self.MarketOrder("SPY", 1) 
       else:
           
           if self.RSI1.Current.Value > 70:
               self.Liquidate()

I tried outputting the contents of the numpy array: self.change_from_close to a csv, but it was unsuccessful. Is there a way to see the contents of a numpy array from within the backtest? Normally I would just use Jupyter and print the output.

Louis Szeto You seem pretty knowledgeable about the price aspect of equities. Do you have any idea what I am doing wrong?

Hi Christopher

I wish the following could help you. A few points to mentions:

1. You need to use Symbol object to call history data, not string: “SPY” → self.Symbol('SPY')
    
2. You have to call History or slice data in OnData function in order to keep it updating
    
3. History function return a dataframe, not a dictionary, you should check if it is empty first, then use .close.unstack("symbol").iloc[-1] to get the data; or my suggestion, use data slice function
    
4. Since it is not in your portfolio at beginning, you should call Security object rather than PortfolioManager
    
5. You have to set up your RSI indicator at initialize to use it

A bit of my personal suggestion, we would not use absolute value change in price usually as it is not stationary, try to use percentage change or log price change to eliminate the compound effect ;)

Wow, thanks so much for your help Louis Szeto ! I really appreciate your input, as the learning curve with QC is quite steep, even for someone experienced with Python.