Update robot forecast

[seabbs]

• Updates Rt forecast for deaths and cases
• Updates convolution forecast based on cases for deaths
• Adds package load checks when needed.

Not staggeringly convinced by any of these forecasts - in particular the German deaths forecasts. No obvious errors though so I think its just model misspecification to whatever is happening in the underlying dynamics.

Do we need additional meta-information for the new model when it comes to submission time?

Rt model

Convolution model

[nikosbosse]

hm interesting - the German deaths really look a bit off. Any ideas why?

[seabbs]

no not sure. as I said it doesn't look like there is a bug it might just be a disconnect between cases and deaths.