QuantConnect Community

see image enclosed

In this algorithm, the trading signals depend on the EMA cross and three daily indicators (RSI, MACD and Momentum). Since the start and end date are set to have the difference of only one day, the daily indicators will not be ready (in other words, there is not sufficient data to calculate them).
There are two options to fix this:

1. Set an earlier StartDate and in OnFifteenMinuteData handler, check whether all indicators are ready before use them to make trading decisions://In Initialize, e.g.:
SetStartDate(2016, 10, 1);

public void OnFifteenMinuteData(
object sender, TradeBar bar)
{
// update our indicators
Fast.Update(Time, bar.Open);
Slow.Update(Time, bar.Close);

var isReady = Fast.IsReady && Slow.IsReady &&
DailyMacd.IsReady && DailyRSI.IsReady &&
DailyMomentum.IsReady;

if(!isReady) return;

/* Continue with trading logic */
}

2. Set warmup with the number of days we need to get the indicators ready (please checkout the docs under the Warming Your Algorithm subsection:

   // In Initialize

// Warm up 30 days of data.
SetWarmup(TimeSpan.FromDays(30));

We should not use the first option for live algorithms, since it would take several days for a deployed algorithm to make its first trade.

Thanks for the reply. I adjusted the RSI to minute instead of daily, it still doesn't execute properly. It always moves three bars back from the screenshot enclosed.

And I move the date to 1,1,2016, which should be more sufficient, no? Thanks again

It impossible to ready the values from those images. 
You need to compare the input (price data) and output (indicator values) information from both platforms to find the origin of those trading signals difference.

 

Thanks for the reply.  If this were to go live, what do I get?  Would it be consistent? 

In live mode, we get the same behavior, since the indicator computation does not depend on the mode (live/backtesting). In other words, yes, the indicator outputs would be consistent.