How to use indicator computed in a dataframe to enter trade

Hi Ang

I think this link has the definition of SCTR. So I'll use QC indicator to create one. Owing to bunches of indicator usage, I've created a new class object for easy and clear manipulation. Using indicators and rolling window will save you lots of time/space from memory leaks in pandas dataframe (repeated call history). 

I attached backtest with some basics codes only. You can modify by your need.

Cheers
Louis

Amendment:

Also need to include the update method for renewal of PPO's rolling window:

	def OnData(self, data):
        
        SCTR = {}
        
        for symboldata in self.Data.values():
        	if not data.ContainsKey(symboldata.symbol): continue
        	symboldata.Update()
            if symboldata.IsReady():
            	pass	# your action

class SymbolData:
    
    def __init__(self, algo, symbol):
        
        self.symbol = symbol
        
        self.EMA200 = algo.EMA(symbol, 200, Resolution.Daily)
        self.EMA50 = algo.EMA(symbol, 50, Resolution.Daily)
        self.ROC125 = algo.ROC(symbol, 125, Resolution.Daily)
        self.ROC20 = algo.ROC(symbol, 20, Resolution.Daily)
        self.PPO = algo.PPO(symbol, 12, 26, MovingAverageType.Exponential, Resolution.Daily)
        self.RSI14 = algo.RSI(symbol, 14, MovingAverageType.Simple, Resolution.Daily)
        
        self.PPOWindow = RollingWindow[float](3)
        
        history = algo.History(symbol, 200, Resolution.Daily)
        for index, bar in history.loc[symbol].iterrows():
            self.EMA200.Update(index, bar.close)
            self.EMA50.Update(index, bar.close)
            self.ROC125.Update(index, bar.close)
            self.ROC20.Update(index, bar.close)
            self.PPO.Update(index, bar.close)
            self.RSI14.Update(index, bar.close)
            
            self.PPOWindow.Add(self.PPO.Current.Value)
            
    def Update(self):
        self.PPOWindow.Add(self.PPO.Current.Value)
        
    def IsReady(self):
        return self.EMA200.IsReady and self.EMA50.IsReady \
                    and self.ROC125.IsReady and self.ROC20.IsReady\
                    and self.PPO.IsReady and self.RSI14.IsReady and self.PPOWindow.IsReady
                    
    def SCTR(self):
        return self.EMA200.Current.Value*0.3 + self.EMA50.Current.Value*0.15 + self.ROC125.Current.Value*0.3 + self.ROC20.Current.Value*0.15 \
                + (self.PPOWindow[0] - self.PPOWindow[2])/(3*self.PPOWindow[2]) *0.05 + self.RSI14.Current.Value*0.05
        

Hi Ang Jin Wei and Louis,
Thank you, Louis, for your response the logic for SCTR. Moving forward, we have implemented the above SCTR logic into a custom python indicator. Refer to the attached backtest for reference.
Best,
Varad Kabade

Hi Ang Jin Wei, Louis, and Varad,

I recently had a client have me work on an algo with this indicator. He mentioned this post, so I was looking at the codes provided to save some time implementing it. I noticed a few things that need updated for this indicator:
 

• The indicator score uses the percent above/below the emas, not the actual ema values. Currently the codes are just using the raw ema scores.
• The short term PPO histogram slope is the most complicated component to calculate. 
  • First off, in QC to get the PPO histogram, we need to calculate the PPO signal. Here's a link on the PPO from Stock Charts. The built-in PPO indicator only gives us the PPO. We have to create another EMA on the PPO to get the PPO signal line. Once we have that we can use indicator extensions or manually take the difference between the two to get the PPO histogram. 
  • Once we have the PPO histogram and add to the rolling window, it's still a custom calculation to get the component score to use. Here's how Stock Charts' describes it here:
    • “For the PPO-Histogram, if its slope is greater than +1 (i.e., +45 degrees), 5 points (5% of 100) is contributed to the total indicator score. If the PPO-Histogram's slope is less than -45 degrees, 0 points are contributed to the score. Otherwise, 5% of ((Slope + 1) x 50) is contributed.”

To see my implementation of the custom indicator, see the “stockcharts_technical_rank” file in the attached backtest. I just added my implementation as a separate file to Varad's backtest. NOTE: it's not using my implementation. I like to use a standard structure where I have a SymbolData class that holds all indicators. It was significantly different, so I thought I'd just share the SCTR indicator class. Hope it's helpful.

Best,

Aaron