Additional comments from Seb

[eahowerton]

• Some of the plots produce warnings (from cowplot::get_legend). Would it be worth trying to get rid of these?

• In the FluSight examples you exclude the baseline model. Is there a particular rationale for this? I thought it was good practice to include this for regularisation purposes.

• You're saying that "we evaluate the performance of these ensembles using scoring metrics that measure..." but then load the scores from a file and don't actually show how you're doing the evaluation. I know you're trying to not show too much detail but perhaps it would be good to have code for the evaluation somewhere for interested readers to check/see?

[lshandross]

The second paragraph of Section 5 has been updated in Commit 364aa4e to more directly specify that any code not shown in the manuscript can be found elsewhere in the linked repository. To me, this feels sufficient to address concerns (3)