great investigative survey about all the different models around covid.

Really split on this one; I consider the article credible in how good a job it does in presenting the depth of scientific opinions on multiple models of continued covid deaths. However, this analysis avoids what is becoming the largest elephant in the room in covid debates: how are we defining a "covid death"? The article assumes that this is settled science, but in reality it is widely acknowledged that a large percent of "covid deaths" are never actually tested if they meet certain subjective criteria, and there are also many examples of people actually dying from other causes WHILE they also have tested positive for covid. In both examples we are opening the possibility of deaths themselves being much more overrepresented than is scientifically accurate; and in that case, all the statistical models presented here are therefore highly suspect as they are modeling possibilities for a stat (covid deaths) that are in themselves not historically accurate. I'd be more wiling to give this article credit if several of the models presented actually reduced the total number of deaths (alongside others that are even higher) instead of just started from the assumption that existing covid death totals are accurate.

The article is well-sourced and does a great job of presenting various easy-to-read Covid-19 models. I agree with reviewer Dave S. that this is a hard one to review because although there is a variety of models being presented, there is no variety in how the models are determining what qualifies as a Covid-19 death. I appreciate the presentation and understand that the authors are simply pulling as many available models as possible, but among the listed assumptions for each graph, it should state how they define a Covid-19 death.

An excellent visualization of the research collected and models generated by different research organizations.

This provides graphs, but no raw data, no model computations, and no identified variables, or the range of values those variables can contain. This is mostly narrative, nothing more.