Why are there two estimates of R issued today, and why are they so different?

The first estimate, the Government estimate, comes from SAGE (the Scientific Advisory Group for Emergencies). These come from SPI-M (the modelling subcommittee of SAGE) where a variety of expert teams from @LSHTM LSHTM, Imperial and others estimate R

In order to get a reliable estimate, you have to use data. In order to get *all* epidemiological data, you require deaths. And these take time - a few days for people to become infected, and a couple of weeks in hospital.

It's described here https://www.gov.uk/guidance/the-r-number-in-the-uk - includes the *significant* caveat that SAGE's estimate of R is the situation 'over the last few weeks': "These estimates do not yet fully reflect any very recent changes in transmission due to, for example, the reopening of schools"

The reason for this is that deaths take a time to show up. I am not sure why deaths rather than cases are needed. I assume death data are much more reliable than case data (particularly when there are limited tests available).

I am not sure why SAGE needs to have death data before releasing an estimate of R. It may be that all modelling groups are required to submit their estimates. I am not sure about this and will look through the SAGE minutes when I have time.

The other estimate of R comes from the REACT study at @MRC_Outbreak https://www.imperial.ac.uk/media/imperial-college/institute-of-global-health-innovation/public/Resurgence-of-SARS-CoV-2-in-England--detection-by-community-antigen-surveillance.pdf - they use swab data and does not need to wait for death data.

My view is the REACT estimate of R being 1.7 is more current and I expect it to be more up-to-date given the SAGE caveats. But I will check.