The Coronavirus, not the beer

[thethe]

i'd love to know what the rate of transmission was

can you tell what it was on any particular day

if you can i would have to change my assessment of you completely

tell me what it was on 3/21 and what it was on 3/31

and please show your work

The change in the reported cases is a proxy for that.

The rate of cases reported dropped SIGNIFICANTLY. That indicates there was less uninfected to transmit the virus to.

The underlying properties of the virus doesn't change. Just the population to spread it to.

My work is a simple spreadsheet that shows the rate of increase in 2 and 3 day cases.

[thethe]

Think that's three revisions now

It will continue.

But please, continue to listen to the 'experts'.

[nsacpi]

The change in the reported cases is a proxy for that.

The rate of cases reported dropped SIGNIFICANTLY. That indicates there was less uninfected to transmit the virus to.

The underlying properties of the virus doesn't change. Just the population to spread it to.

My work is a simple spreadsheet that shows the rate of increase in 2 and 3 day cases.

so what was the transmission rate on 3/21 and on 3/31 and how did you get those numbers

[thethe]

so what was the transmission rate on 3/21 and on 3/31 and how did you get those numbers

You compare the most recent period set with the current period set

Current Period 2 or 3 day growth / Prior period 2 or 3 day growth - 1


I gave the rate of transmission on 3/21 already. For 3/31:

2 day (4/1) - 32%
3 day (3/30) - 23%

[nsacpi]

You compare the most recent period set with the current period set

Current Period 2 or 3 day growth / Prior period 2 or 3 day growth - 1


I gave the rate of transmission on 3/21 already. For 3/31:

2 day (4/1) - 32%
3 day (3/30) - 23%

my understanding of transmission rate is how many people an infected person infects...is your definition different

[thethe]

my understanding of transmission rate is how many people an infected person infects...is your definition different

In a hypothetical world where you are the only person alive and you contract a massively contagious virus its irrelevant how contagious it is. You will not transmit it to anyone.

So the virus has its properties and then the real transmission rate is based on the enviornment.

[nsacpi]

In a hypothetical world where you are the only person alive and you contract a massively contagious virus its irrelevant how contagious it is. You will not transmit it to anyone.

So the virus has its properties and then the real transmission rate is based on the enviornment.

so how do you define transmission rate

[thethe]

so how do you define transmission rate

The amount of people that are being infected in a period divided by the amount of people infected in the prior period (minus 1 for a percentage).

The whole point it to understand how high is the peak. If there are less people contracting it proportionally it means there is less people to infect. Of course the aggreagte number is getting higher but once you get to 0 you will fall off a cliff because there just aren't any uninfected people along normal commuting routes to trasnmit the virus.

[nsacpi]

The amount of people that are being infected in a period divided by the amount of people infected in the prior period (minus 1 for a percentage).

Ok. That's a reasonable definition.

To calculate it we need the number of people infected in the two periods being compared. Or proxies for those numbers.

So what are your proxies for number of people infected in the two periods.

[thethe]

Ok. That's a reasonable definition.

To calculate it we need the number of people infected in the two periods being compared. Or proxies for those numbers.

So what are your proxies for number of people infected in the two periods.

The confirmed cases. I stated in one of my previous post that this is a horrible proxy but its the best one we have now.

[nsacpi]

The confirmed cases. I stated in one of my previous post that this is a horrible proxy but its the best one we have now.

I think this is where you can get very weird results if you look at data when the infection is first detected. Testing is very low at the start. You get 1 case that first day. And obviously that number is meaningless. By day 100, testing has been ramped up substantially and by then number of confirmed cases is a much better proxy.

There are additional issues. One is the number of cumulative cases needs to be adjusted for people who are no longer contagious.

So I think your proxy is not bad for the mature phase of the epidemic.

I would look to refine it a little. New cases divided by cases reported in the prior 15 days. That should eliminate those who tested positive a while back and are presumably no longer contagious.

I don't think your method is reliable as a way to compare March 21 to March 31, simply because the relationship between number infected and the proxy you are using for number infected changed substantially over that period due to the expansion of testing.

The approach you are taking is not bad. But I think you need to be careful drawing inferences over periods when the relationship between number infected and your proxy was likely very unstable.

[thethe]

I think this is where you can get very weird results if you look at data when the infection is first detected. Testing is very low at the start. You get 1 case that first day. And obviously that number is meaningless. By day 100, testing has been ramped up substantially and by then number of confirmed cases is a much better proxy.

There are additional issues. One is the number of cumulative cases needs to be adjusted for people who are no longer contagious.

So I think your proxy is not bad for the mature phase of the epidemic.

I would look to refine it a little. New cases divided by cases reported in the prior 15 days. That should eliminate those who tested positive a while back and are presumably no longer contagious.

I don't think your method is reliable as a way to compare March 21 to March 31, simply because the relationship between number infected and the proxy you are using for number infected changed substantially over that period due to the expansion of testing.

The approach you are taking is not bad. But I think you need to be careful drawing inferences over periods when the relationship between number infected and your proxy was likely very unstable.

During those specific periods of time the CCP Virus knowledge was widespread. Therefore, the testing was happening for people who had symptoms. Of course testing has ramped up but it does not account for the massive drop in rate of confirmed cases when it seems to have reached a critical mass point.

No analysis is a sure thing. That is why these models change seemingly by the day now. But the downward revision leads more credence to the idea that we hit peak much earlier than anticipated.

[nsacpi]

During those specific periods of time the CCP Virus knowledge was widespread. .

I agree with this. But the prevalence of testing changed a lot from March 21 to 31. That's the problem. The proxy obviously understates how many people were infected. But it understates how may were infected on March 21 by a lot more. And that will make it look like the transmission rate dropped by a lot more than it actually did from March 21 to March 31.

[thethe]

I agree with this. But the prevalence of testing changed a lot from March 21 to 31. That's the problem. The proxy obviously understates how many people were infected. But it understates how may were infected on March 21 by a lot more.

I guess the question is then were there people who reported symptoms to hopsitals or PHP that could not get a test during the mid march period. If that is the case then I agree with what you are trying to say.

Even still, while that may impact the rate of change the overall trend line would not change.

[thethe]

Who knows if this is even real but all of our media companies should be investigated for why they are pushing CCP propaganda.

[nsacpi]

Even still, while that may impact the rate of change the overall trend line would not change.

Here I think you are assuming something without any factual basis. We just don't know whether the transmission rate changed from March 21 to 31.

Some very smart statistician might find ways to adjust for the change in testing prevalence. But I haven't seen anything like that.

[thethe]

Here I think you are assuming something without any factual basis. We just don't know whether the transmission rate changed from March 21 to 31.

Some very smart statistician might find ways to adjust for the change in testing prevalence. But I haven't seen anything like that.

Is there evidence that people with symptoms couldn't get testing for the CCP virus in mid march?

[nsacpi]

Is there evidence that people with symptoms couldn't get testing for the CCP virus in mid march?

Yes. It was a common complaint. If you played in the NBA you could get a test. For everyone else that wasn't the case.

[thethe]

Yes. It was a common complaint. If you played in the NBA you could get a test. For everyone else that wasn't the case.

So if your contention is that earlier in the month less confirmed cases existed because of a lack of testing then that would imply the rate of increase later on should be even LESS

[nsacpi]

So if your contention is that earlier in the month less confirmed cases existed because of a lack of testing then that would imply the rate of increase later on should be even LESS

during the period when testing is being ramped up your proxy would be biased in the direction of exaggerating the transmission risk...and once testing availability stabilizes the bias eliminated

so you want to be careful about drawing inferences that involve that period when testing availability changed

it is probably worth going back and looking at the number of tests being performed each day and seeing when that was rising fast and at what point it stabilized