Hi it's Michael here from Norwich
Hi everybody! If you have questions as we go, please do type in here. There will be time at the end for questions too. Thanks!
Not quite true.
Want to come in Chris?
@Paul happy to wait until the end.
Roger!
SAGE forecasts are produced using a principled ensemble summarisation method developed by DSTL (see "CrystalCast", Bowman et al.). They are intended to allow for model uncertainty -- different groups will assume different values for inherently unmeasurable quantities.
...and such quantities are not always easy to estimate, even with Kalman smoothing.
So you don't believe that a randomised study design is a sensible thing to do to measure prevalence? Instead, preferring to trust data from the TTI scheme which is designed for control, not for surveillance?
[Thomas] A lot has been overly secret; but there are some better plots just very recently released at: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/932317/201031_SPI-M-O_medium_term_projections_explainer.pdf
Chris / Thomas - @David Leslie will invite you both to comment at the end. Is that ok?
@Paul sure.
Thanks James
If anyone has any questions, please stick them here or indicate here that you’d like to ask one
[Michael] Thank you very much for your modeling efforts since the spring and for emphasising the need for calibrating the models and bringing in inter-disciplinary expertise to help.
[Thomas] I’ll come in as well
[Kate] this is an interesting article: https://theconversation.com/next-slide-please-data-visualisation-expert-on-whats-wrong-with-the-uk-governments-coronavirus-charts-149329?utm_source=Nature+Briefing&utm_campaign=fc6bf4b9de-briefing-dy-20201104&utm_medium=email&utm_term=0_c9dfd39373-fc6bf4b9de-45681514
[Dr Kieron] Thanks all from Geneva. :-) [Email redacted]