Questions/comments after edits

[jgabry]

I pushed a bunch of edits yesterday. Here are two things I didn't change but maybe we should?

• Since the algorithm proceeds backwards in time, should Figure 1 be flipped? Otherwise it needs to be read from right to left, correct? Or am I misunderstanding?

• It looks like for the case studies we show predictive intervals for the Lake Huron example (Figure 5) but not for the cherry blossom example (Figure 8). That is, it looks like in Figure 8 the uncertainty interval is just the uncertainty in the mean rather than in the prediction. Should we change that to match Figure 5?

[paul-buerkner]

Thanks Jonah!

1. You are right. Indeed, we should perhaps reverse the order of the x-axis to reflect the flow of the algorithm.

2. You are right. I did show the mean in Figure 8 instead of the PPs as the CIs of the latter where soo wide that the trend was a little bit hard to see. In the interest of consistency, we may want to also plot the means in Figure 5 as I don't think Figure 8 looks that great with PPs (although I generally prefer PPs by a lot!). I am open to other suggestions of course!