Hey!
I've been finishing the bootcamp called SWPCM model, and thought it would be a good idea to mess around with the algorithm, to get a better grasp on the Lean framework. First, i worked around a bug in the OnSecuritiesChanged part which worked wonderfully. Futhermore, i thought it would be interresting to rebalance the portfolio every 4th week, instead of everyday, as i saw a lot of small transactions. I've tried to implement the feature in the following code.
from datetime import timedelta
from QuantConnect.Data.UniverseSelection import *
from Selection.FundamentalUniverseSelectionModel import FundamentalUniverseSelectionModel
class SectorBalancedPortfolioConstruction(QCAlgorithm):
def Initialize(self):
self.SetStartDate(2015, 12, 28)
self.SetEndDate(2020, 12, 31)
self.SetCash(100000)
self.UniverseSettings.Resolution = Resolution.Hour
self.SetUniverseSelection(MyUniverseSelectionModel())
self.SetAlpha(ConstantAlphaModel(InsightType.Price, InsightDirection.Up, timedelta(1), 0.025, None))
self.SetPortfolioConstruction(MySectorWeightingPortfolioConstructionModel(timedelta(weeks=4)))
self.SetExecution(ImmediateExecutionModel())
self.Settings.RebalancePortfolioOnInsightChanges = False
self.Settings.RebalancePortfolioOnSecurityChanges = False
class MyUniverseSelectionModel(FundamentalUniverseSelectionModel):
def __init__(self):
super().__init__(True, None, None)
def SelectCoarse(self, algorithm, coarse):
filtered = [x for x in coarse if x.HasFundamentalData and x.Price > 0]
sortedByDollarVolume = sorted(filtered, key=lambda x: x.DollarVolume, reverse=True)
return [x.Symbol for x in sortedByDollarVolume][:100]
def SelectFine(self, algorithm, fine):
filtered = [f for f in fine if f.AssetClassification.MorningstarSectorCode == MorningstarSectorCode.Technology]
self.technology = sorted(filtered, key=lambda f: f.MarketCap, reverse=True)[:3]
filtered = [f for f in fine if f.AssetClassification.MorningstarSectorCode == MorningstarSectorCode.FinancialServices]
self.financialServices = sorted(filtered, key=lambda f: f.MarketCap, reverse=True)[:2]
filtered = [f for f in fine if f.AssetClassification.MorningstarSectorCode == MorningstarSectorCode.ConsumerDefensive]
self.consumerDefensive = sorted(filtered, key=lambda f: f.MarketCap, reverse=True)[:1]
return [x.Symbol for x in self.technology + self.financialServices + self.consumerDefensive]
class MySectorWeightingPortfolioConstructionModel(EqualWeightingPortfolioConstructionModel):
def __init__(self, rebalance = Resolution.Daily):
super().__init__()
self.symbolBySectorCode = dict()
self.result = dict()
def DetermineTargetPercent(self, activeInsights):
#1. Set the self.sectorBuyingPower before by dividing one by the length of self.symbolBySectorCode
self.sectorBuyingPower = 1/len(self.symbolBySectorCode)
for sector, symbols in self.symbolBySectorCode.items():
#2. Search for the active insights in this sector. Save the variable self.insightsInSector
self.insightsInSector = [insight for insight in activeInsights if insight.Symbol in symbols]
#3. Divide the self.sectorBuyingPower by the length of self.insightsInSector to calculate the variable percent
# The percent is the weight we'll assign the direction of the insight
self.percent = self.sectorBuyingPower / len(self.insightsInSector)
#4. For each insight in self.insightsInSector, assign each insight an allocation.
# The allocation is calculated by multiplying the insight direction by the self.percent
for insight in self.insightsInSector:
self.result[insight] = insight.Direction * self.percent
return self.result
def OnSecuritiesChanged(self, algorithm, changes):
for security in changes.AddedSecurities:
algorithm.Log(f"Remove {security.Symbol}")
if security.Fundamentals is None:
raise AttributeError(f"{security.Symbol} has fundamentals")
sectorCode = security.Fundamentals.AssetClassification.MorningstarSectorCode
if sectorCode not in self.symbolBySectorCode:
self.symbolBySectorCode[sectorCode] = list()
self.symbolBySectorCode[sectorCode].append(security.Symbol)
sectorsToRemove = []
for security in changes.RemovedSecurities:
algorithm.Log(f"Remove {security.Symbol}")
if security.Fundamentals:
algorithm.Log(f"{security.Symbol} has fundamentals")
for sectorCode, symbols in self.symbolBySectorCode.items():
if security.Symbol in symbols:
symbols.remove(security.Symbol)
if not symbols:
sectorsToRemove.append(sectorCode)
for sectorCode in sectorsToRemove:
self.symbolBySectorCode.pop(sectorCode)
super().OnSecuritiesChanged(algorithm, changes)
I get a runtime keyerror when i try to run it. It breaks in the middle of the backtest. I've been trying to search online and fix it, but i can't seem to do it..... The problem seems to be related to the algorithm not being able to fetch certain historical data, but i dont really know.
Any solutions will be gladly accepted!
Best.
Lucas
(sorry for bad english, not my first language)
Vladimir
Lucas Bjerre Rasmussen,
I was able to run your algorithm by replacing line 46
from
self.sectorBuyingPower = 1/len(self.symbolBySectorCode)
to
self.sectorBuyingPower = 1/len(self.symbolBySectorCode) if len(self.symbolBySectorCode) > 0 else 0
Derek Melchin
Hi Lucas,
Two issues arise when executing the code above.
First, we need to ensure we only remove a key from `self.symbolBySectorCode` if the key exists in the dictionary. We can accomplish this by replacing
for sectorCode in sectorsToRemove: self.symbolBySectorCode.pop(sectorCode)
with
for sectorCode in set(sectorsToRemove): if sectorCode in self.symbolBySectorCode: self.symbolBySectorCode.pop(sectorCode, None)
Additionally, the algorithm above receives margin calls. We can avoid these by increasing both the cash buffer and the rebalancing frequency.
self.Settings.FreePortfolioValuePercentage = 0.2 self.SetPortfolioConstruction(MySectorWeightingPortfolioConstructionModel(timedelta(weeks=1)))
See the attached backtest for reference.
Best,
Derek Melchin
The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by QuantConnect. In addition, the material offers no opinion with respect to the suitability of any security or specific investment. QuantConnect makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances. All investments involve risk, including loss of principal. You should consult with an investment professional before making any investment decisions.
Lucas
The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by QuantConnect. In addition, the material offers no opinion with respect to the suitability of any security or specific investment. QuantConnect makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances. All investments involve risk, including loss of principal. You should consult with an investment professional before making any investment decisions.
To unlock posting to the community forums please complete at least 30% of Boot Camp.
You can continue your Boot Camp training progress from the terminal. We hope to see you in the community soon!