New University of Washington projections accounting for omicron

oragator

Joined: Dec 10, 2011

Posts: 8,005

User Profile Ignore User Stop Ignoring

How long do you want to ignore this user?

24 hours One week Permanently Cancel

Both surreal and somewhat comforting at the same time.
They have us getting to around 2.8 million cases a day, but relatively low hospitalization and death rates. But around 150k more deaths by April 1.

https://covid19.healthdata.org/united-states-of-america?view=infections-testing&tab=trend&test=infections

1

DCAggie13y

Joined: Jul 19, 2001

Posts: 6,103

User Profile

Private Message

Ignore User Stop Ignoring

How long do you want to ignore this user?

24 hours One week Permanently Cancel

In reply to oragator • 7:31p, 12/22/21

AG

This model has been horribly wrong for a while now. It used to be decent but it appears they were pressured into making it spit out worst case data about a year ago. They even adjusted their death numbers to account for uncounted deaths and make the death totals higher.

I looked at their funders and it looks like they received some funding from the usual suspects so that may be why they adjusted everything to exaggerate their forecasts.

Bottom line. It's a terrible model.

19

1 edit

oragator

Joined: Dec 10, 2011

Posts: 8,005

User Profile Ignore User Stop Ignoring

How long do you want to ignore this user?

24 hours One week Permanently Cancel

In reply to DCAggie13y • 7:50p, 12/22/21

Their last projection from November actually understated the current death toll by about 10k.
But I agree it shouldn't be treated as gospel and could be off, even substantially. Even if they double counted the number of actual cases though it is going to be close to 1.5 million a day.

We will see, it's one data point, there will be others

5

1 edit

amercer

Joined: Apr 25, 2003

Posts: 29,002

User Profile Ignore User Stop Ignoring

How long do you want to ignore this user?

24 hours One week Permanently Cancel

7:54p, 12/22/21

AG

The pessimists have been way closer to correct than the optimists for the last 18 months. But we will see.

6

El Chupacabra

Joined: Feb 13, 2007

Posts: 7,705

User Profile Ignore User Stop Ignoring

How long do you want to ignore this user?

24 hours One week Permanently Cancel

In reply to amercer • 8:03p, 12/22/21

amercer said:
The pessimists have been way closer to correct than the optimists for the last 18 months. But we will see.

millions dead in the US in the first 6 months alone!

17

DCAggie13y

Joined: Jul 19, 2001

Posts: 6,103

User Profile

Private Message

Ignore User Stop Ignoring

How long do you want to ignore this user?

24 hours One week Permanently Cancel

In reply to oragator • 10:16p, 12/22/21

AG

oragator said:
Their last projection from November actually understated the current death toll by about 10k.
But I agree it shouldn't be treated as gospel and could be off, even substantially. Even if they double counted the number of actual cases though it is going to be close to 1.5 million a day.

We will see, it's one data point, there will be others

It doesn't surprise me that their forecast was inaccurate for November. Their model has been pretty bad for a while now.

1

Clown_World

Joined: Aug 7, 2006

Posts: 5,161

User Profile Ignore User Stop Ignoring

How long do you want to ignore this user?

24 hours One week Permanently Cancel

11:11p, 12/22/21

Academics have never projected anything that ended up totally wrong.

Like ever.

May as well cancel the bowl games and sit around and get fat while we wait for the next injection.

1

Gordo14

Joined: Mar 2, 2012

Posts: 12,000

User Profile

Private Message

Ignore User Stop Ignoring

How long do you want to ignore this user?

24 hours One week Permanently Cancel

In reply to El Chupacabra • 11:42p, 12/22/21

El Chupacabra said:
amercer said:
The pessimists have been way closer to correct than the optimists for the last 18 months. But we will see.

millions dead in the US in the first 6 months alone!

Literally nobody ever said that but sure.

Most models had deaths bound between 500k and 1.2MM for a 1 year prediction. Deaths were I think in the upper 600k, but a large reason for the number being on the low end was the speed and effectiveness of the vaccine.

I think this model is flawed in that there's still a lot of uncertainty on hospitalization and death data. As far as infections are concerned I think the model is pretty accurate and reasonable.

2