Reviewer 1 Comment:
Paragraph on line 295: or maybe single-country models perform better because multi-country models are fitted to data from different ongoing epidemics, and their shared or hierarchical parameters may simply be not as well calibrated for a specific country. i.e., the advantage may simply be conferred by the choice of training data, not by bespoke approaches taken by modellers. I do not think you have shown evidence that modellers of single-country models have put in more bespoke features in their models.
Reviewer 1 Comment:
Paragraph on line 295: or maybe single-country models perform better because multi-country models are fitted to data from different ongoing epidemics, and their shared or hierarchical parameters may simply be not as well calibrated for a specific country. i.e., the advantage may simply be conferred by the choice of training data, not by bespoke approaches taken by modellers. I do not think you have shown evidence that modellers of single-country models have put in more bespoke features in their models.