Built a benchmark for simple buy and sell

Based on the information provided in the extracts, it seems like you want to create a benchmark that simulates buying and holding a portfolio of stocks, while also running a machine learning model for actual trading.

To create a benchmark in QuantConnect, you can use the SetBenchmark function. However, this function typically sets a single security (like SPY) as the benchmark. It doesn't directly support a custom portfolio as a benchmark.

One possible workaround could be to create a separate algorithm that buys and holds your desired portfolio, and then use the results of this algorithm as a benchmark for your main algorithm. This separate algorithm would take into account dividends, splits, etc. However, this would require running two separate backtests and comparing the results manually.

As for the machine learning part, QuantConnect does support machine learning algorithms. You can plug in your ML model, such as Random Forest, with the appropriate hyper-parameters. The algorithm is a work-in-progress and the QuantConnect team is continuously working to improve it. You can adapt it to handle multiple symbols and use daily data consolidated from minute data.

Here is an example of how you can use the SetBenchmark function:

self.benchmarkTicker = 'SPY' 
self.SetBenchmark(self.benchmarkTicker)

And here is an example of how you can use a machine learning model in your algorithm:

# Assuming you have a trained model 'model'
prediction = model.predict(data)
if prediction == 1:    # The model predicts the price will go up
    self.SetHoldings("SPY", 1)
elif prediction == -1: # The model predicts the price will go down
    self.SetHoldings("SPY", -1)

Please note that these are just basic examples. You would need to adapt them to fit your specific needs and trading strategy.

For more information, you can refer to the QuantConnect documentation and the QuantConnect forum for discussions related to your query.

Great thanks for your reply. 

You're welcome, Krish! If you have any more questions or need further clarification, feel free to ask. Happy coding!

Hi Mia,

I want to know if can I run two algorithms parallelly. 

class CustomTickerAlgorithm(QCAlgorithm):

    def Initialize(self):
        self.SetStartDate(2021, 5, 13)
        self.SetEndDate(2021,7,13)
        self.SetCash(100000)
        self.AddEquity("SPY", Resolution.Daily)
        self.fakeqc = FakeQCModel()
        
    def OnData(self, data: Slice):
        if not self.Portfolio.Invested:
            self.SetHoldings("SPY", 0.33)
        else:
            self.Log("ankhsdgjua")

class FakeQCModel(QCAlgorithm):
    def Initialize(self) -> None:
        pass

    def OnData(self, slice: Slice) -> None:
        return super().OnData(slice)

I want to build two portfolios, CustomTickerAlgorithm is my main algorithm where I write my Machine learning algorithm, and based on ML it will buy and sell stocks and build real real-time Portfolio. 

in my FakeQCModel there will be a simple buy and hold Portfolio which can track some assumption or hypothesis.

from my main class, I want to take a value of fake portfolio for regular interval. 

is it possible?

I have tried to run but it does not call Ondata of FakeQCModel, also tried to implement with Alpha model but did not work.

Based on the information provided in the extracts, it's not possible to run two algorithms in parallel within the same brokerage account. This is because QuantConnect/Lean only manages one portfolio per account. To place orders, it needs to know the portfolio state for buying power calculation. Consequently, it would need to fetch information from each account and save it in different portfolios with different states/buying power.

However, you can run multiple strategies within a single algorithm. You can call different strategies within the OnData method of your algorithm. Here's an example:

def OnData(self, data):
    self.CallStrategy1(data)
    self.CallStrategy2(data)

def CallStrategy1(self, data):
    # Your strategy 1 logic here

def CallStrategy2(self, data):
    # Your strategy 2 logic here

In your case, you could have one strategy for your machine learning model and another strategy for your buy and hold portfolio. You would need to manage the cash allocation between the two strategies manually.

For more information, you can refer to the QuantConnect documentation and the QuantConnect forum for discussions related to your query.

Thank you for your reply.
let me explain my implementation or code, I'm new here so some terminology or concept may I miss

1. My machine learning model will trade daily based on prediction, and based on it one portfolio will be managed.
2. in my fake FakeQCModel there will be just one trade like I run self.SetHolding("SPY", 1). and leave it for my end algorithm date. 
3. so now I have two portfolios which have 2 different types of trading value one which I trade regularly (buy/sell) and got some profit/loss percentage. and another have only one trade which I bought at the starting of my backtest. now I will compare both what percentage loss/profit will I get If I use ML model.
   
   this is what I'm doing

it runs well if I build 2 different algorithms and run both manually. but I want to merge both and my main algorithm call it self at the end of backtest, and get profit/loss from fakeQcModel Portfolio.

Hello Krish Nandan das and Sukhvinder Pal Singh 
This is not possible sorry.
We cannot create multiple subclasses from the QCAlgorithm class.