QuantConnect Community

Hi Jesse,

if you ‘restart’ the algorithm, will it also restart the warm up period, rendering your algorithm useless for the amount of days it is being warmed up for?

To ensure the algorithm starts trading immediately, call the SetWarmUp method or warm up the algorithm manually with a history request.

Best,
Derek Melchin

Want to invest in QuantConnect as we build the Linux of quant finance? Checkout our Wefunder campaign to join the revolution.

Thank you! 

so if i have SetWarmUp period in initialize, i should be okay? 

Thank you for your time, 

Jesse

Hi Jesse,

If we add SetWarmUp, LEAN winds back the clock and pumps data into the algorithm before the deployment time. If the algorithm uses the warm-up data to correctly warm up any indicators and lookback windows, it should allow the algorithm to start trading immediately. For further assistance, please attach a backtest of the algorithm.

Best,
Derek Melchin

Want to invest in QuantConnect as we build the Linux of quant finance? Checkout our Wefunder campaign to join the revolution.

Thank you so much for taking the time to respond to me. Just to verify because of course it would be very costly if I somehow misunderstood you. 

Take this template for instance, I have the warm up set to 70 days. If I deploy this algorithm and restart it upon deployment, will it go back 70 days automatically and start paper trading correctly immediately?  

class CombinedAlgorithm(QCAlgorithm):

def Initialize(self):

# INITIALIZE

self.SetStartDate(2020, 8, 1) # Set Start Date

self.SetEndDate(2021, 12, 12)

self.SetCash(10000) # Set Strategy Cash

self.symbol = self.AddEquity('TSLA', Resolution.Minute)

self.symbol.SetDataNormalizationMode(DataNormalizationMode.Raw)

# SCHEDULED EVENTS

self.Schedule.On(self.DateRules.WeekStart(self.symbol.Symbol), self.TimeRules.AfterMarketOpen(self.symbol.Symbol), self.WeekStart)

self.Schedule.On(self.DateRules.WeekEnd(self.symbol.Symbol), self.TimeRules.BeforeMarketClose(self.symbol.Symbol), self.WeekEnd)

self.Schedule.On(self.DateRules.EveryDay(self.symbol.Symbol), self.TimeRules.BeforeMarketClose(self.symbol.Symbol, 5), self.ParabolicBarCheck)

self.SetWarmUp(timedelta(days = 70))

# CONSOLIDATED DATA

# ATR

atr_consolidator = TradeBarConsolidator(timedelta(days=1))

self.SubscriptionManager.AddConsolidator(self.symbol.Symbol, atr_consolidator)

atr_consolidator.DataConsolidated += self.ATRDayBar

self.atrBarWindow = RollingWindow[TradeBar](14)

# Daily Bars

dailyConsolidator = TradeBarConsolidator(timedelta(days=1))

self.SubscriptionManager.AddConsolidator(self.symbol.Symbol, dailyConsolidator)

dailyConsolidator.DataConsolidated += self.OnTwoDayBar

self.dayBarWindow = RollingWindow[TradeBar](2)

# Weekly Bars

weeklyConsolidator = TradeBarConsolidator(Calendar.Weekly)

self.SubscriptionManager.AddConsolidator(self.symbol.Symbol, weeklyConsolidator)

weeklyConsolidator.DataConsolidated += self.OnTwoWeekBar

self.weekBarWindow = RollingWindow[TradeBar](2)

# CUSTOM CHART

log_plot = Chart('Custom Chart')

log_plot.AddSeries(Series('Orders',SeriesType.Scatter,0))

log_plot.AddSeries(Series('Symbol Price', SeriesType.Line,0))

self.AddChart(log_plot)

def OnData(self, data):

if self.IsWarmingUp:

return

# ENTRY AND EXIT CODE HERE

Hi Jesse,

It's impossible to know if the algorithm will be immediately ready to trade without the full algorithm. To test if it will be ready, in OnData, add

if not self.IsWarmingUp and not self.atr.IsReady:

    self.Quit("The algorithm won't trade immediately.")

You can extend this guard to check if all your indicators and RollingWindow objects will be ready when the algorithm is done the warm-up period.

Best,
Derek Melchin

Want to invest in QuantConnect as we build the Linux of quant finance? Checkout our Wefunder campaign to join the revolution.

Thank you my friend.