Running a backtest using 'record' from the zipline API provided a plot of the recorded values next to the results of the backtest in Quantopian. In Blueshift, it seems to have no effect. Is it usable?
At present, the use of the "record" API has no impact on the plot output. However, the recorded variables are still available in the 'perfs' dataframe (the second argument in the 'analyze' function, that is called at the end of the algo run).
We are soon releasing a major version upgrade on blueshift (replacing zipline with our inhouse engine that already runs the live version now). After that release, the recorded variables from a "full backtest" will be available in plots in the 'tearsheet' tab. However, the "quick backtest" will continue to exclude the recorded variables from the plot output. We will also introduce a limit on the maximum number of variables that can be recorded.
Please feel free to give a shout in case you need more information.