Louisiana and Arizona will beat New York and New Jersey

Post Reply New Topic

If you go here cnn.com/interactive/2020/health/coronavirus-us-maps-and-cases/ you can click on four things:

1) Total cases
2) Cases per 100 K
3) Total deaths
4) Deaths per 100 K

Well, if you pick "Cases per 100 K", then New York and New Jersey have been long standing 1 and 2, respectively. Well, they are about to be replaced! Soon, the 1 and 2 will be Louisiana and Arizona. I don't know which of them will be 1 which of them will be 2, they are kind of going neck by neck. But seeing how New York and New Jersey double their cases within a year while Louisiana and Arizona do that in a month, they will sure beat them, especially since they are already quite close!

You want to know one more interesting fact? Despite the fact that New York and New Jersey are numbers 1 and 2, they are one of the SAFEST states. You know why? Because what matters is NOT the "cases per 100 K", what matters is "NEW cases per 100 K" -- as in, how likely are you to GET sick (aka become a NEW case). In order to calculate NEW cases per 100 K, you have to divide "cases by 100 K" by the time that it doubles. Now, the average doubling time across the US is 1 month, but in New York and New Jersey its a year. So you have to take New York and New Jersey and divide it by 12. If you don't divide it by 12, you have around 2000, which is A LOT; but if you DO divide it by 12, you get less than 200, which is VERY LITTLE. Now look at all the other states and see which ones have less than 200. Hawaii is the only one! Well, I guess you shouldn't get too excited since New York and New Jersey aren't the only ones that take a year to double. Massachusetts takes 11 months, for example. But still, if you take all the states that do take so long, it won't be very many. So while I am not sure that New York and New Jersey are "the" safest, they are "one of the safest" for sure.

By the way, I suspect that the states that take a year to double are all close to the top. In other words, they had their peaks in the past so now they are no longer at their peak -- while all the OTHER states are. So its not really about "which states are the most dangerous" it is about "which states had their peaks when". New York had one of the earliest peaks, thats why it USED TO BE the most dangerous state but now it became one of the safest states. Same applies to New Jersey. Same applies to a lot of other states at the top.

By the way, Washington got over its peak a lot earlier than New York did. Back when it just started, the two states at the top were New York and Washington. But now Washington is no longer at the top at all. In fact its at the bottom half. New York is still at the top, I guess it took longer for New York to get over its peak -- but if it continues to go at the rate that it does, I am sure it will slide to the bottom just like Washington did.

Interestingly enough, Washington does NOT take a year to double. It doubles every month, which is average across the US. So I guess Washington "forgotten" that it used to be at the top and now it acts like your average state. Perhaps New York will speed up its doubling time too, once it gets to the bottom and "forgets" everything?

I'd like to see these stats adjusted by not only population size, but also average population density, and average wealth per head for each state.

Otherwise the comparisons are pretty meaningless.

There's potentially a very big difference between 1m people all equally spread out and 1m people where 750k live in cities and the remaining 250k are spread out in the countryside.

Average wealth gives you some idea of likely healthcare coverage / quality.

I'm very wary of infection figures as there could be lots of white noise there, either over or under. It's harder to argue with mortality rates. Even so some States might be more likely to classify a death as Covid-19 than others, if there are multiple factors involved, or the testing is patchy.

Last edited by Redd_Kross on 16 Jul 2020, 6:56 pm, edited 1 time in total.

As a resident of Washington state, even here in the more conservative east side, I'd say 99.9 % of people wear masks and practise social distancing. No big secret - - people here are taking the plague seriously.

They already did adjust for population size. They have the section of total cases and they have the separate section per 100 K. In fact in the OP I was focusiong per 100 K.

But its true though that they didn't adjust for other parameters (density and so forth).

Yep apologies, I realised that straight after I hit "submit". Have edited my post now.

Actually I was thinking the opposite. Since it takes a long time for sick people to die, the mortality rates reflect the desease prevalence in the past a lot more than in the present. Thats why -- in terms of mortality per 100 K -- the New York and New Jersey are WAY aheas of Arizona and Louisiana, while in terms of the desease rate the gap is not that big. Why? Thats because Arizona being close to New York desease-wise is a very recent news, which is why death rate didn't have time to catch up to it.

I was talking more about accuracy, whereas I think you're concentrating more on immediacy?

Yes there is always a lag, although sadly with Covid-19 it often isn't very long.

The big difference is with the degree of certainty. Infection figures very often feature anyone who's got a cough, especially where testing is woefully inadequate. Death certificates are normally more accurate, though even then I'm sure there will be some "she had Covid-19 when she got hit by the car, so it's Covid-19" figures in there somewhere.

Both factors are important. In terms of "what is more important, death rate or desease rate" they imply the opposite things. So it is interesting how they play out one against another.

As far as accuracy, one other factor to adjust for is "rate of testing". How many percents of asymptomatic people get tested? How many percents of people that do experience symptoms get tested? And how many percents of each group got positive and negative test results?

I am not saying the testing should be more available (thats a different topic altogether). What I am saying is that -- with the testing as it currently is -- its availability should be included into state by state statistics and adjusted for in the calculation of these numbers.

What this means is that even more sections should be introduced. Some sections should describe numbers without testing rates adjusted for, and the other sections should describe numbers with testing rates adjusted for. Perhaps there are many different formulae to account for testing rate (or not), and the person should be able to select their favorite formulae and get the statistics pertaining to that formula.



Apparently it IS very long -- as evident by New York and New Jersey having such a huge gap when it comes to deaths despite the gap narrowing when it comes to actual numbers. I mean you don't have to look at just Arizona and Louisiana. Even if you look at any of the states that are, say, 700 cases per 100 K (which is a third of New York), their rate of deaths per capita is more like 10% of New York. And since there are quite a few states with 700 cases per 100 K, its not a coincidence.

I haven't looked at too many other states but if I look at California, New Mexico and Colorado, then in the past Colorado had more cases per capita than New Mexico and New Mexico had more cases per capita than California. More recently it got reversed: California has more cases per capita than New Mexico and New Mexico has more cases per capita than Colorado. But, in terms of deaths, it is still the case that Colorado has more deaths per capita than New Mexico and New Mexico has more deaths per capita than California. So that shows that there "is" a noticeable lag between cases per capita and deaths per capita.



Actually I think its the opposite. My former landlord, in Minnesota, had quite obvius symptoms of COVID 19 (he had difficulties breathing etc) but his case wasn't counted because he didn't have positive test result -- and the reason he didn't have it is that he wasn't tested on the first place. He wanted to get tested but the doctors said he wasn't eligible for some reason or other.



There are two issues here:

1) If someone dies from COVID 19, but they weren't diagnosed with it (because they weren't tested), then their death won't be attributted to COVID 19

2) If they *were* diagnosed with COVID 19, but die from something else, their death "will" be attributted to COVID 19 even though it shouldn't have been

I don't know which of these two things would overweight, thats a good question.

By the way, what about George Floyd? As we all know, he had COVID 19, but he died from the cop's knee. Was his death attributted to COVID 19 or not?

So much depends on the critirea the Government set.

When I spoke about time lag, I meant between infection becoming apparent and recovery / partial recovery / death. There are some people who have taken a very long time to recover but they seem to be very much in the minority. That's not the same as plotting initial infection rates because Covid-19 can lie dormant for some time. You can back-work diagnosis dates from death certificates with a reasonable degree of average accuracy but anything prior to that is pot luck.

In general I suspect the "overs and unders" balance each other out with infection / spread, but they're still really vague figures.

Here in the UK we've had next to no testing until very recently, but various shielding, furlough and treatment programmes in place. H&S "right side failure" for safety is to send anyone home who MIGHT be displaying symptoms, but with no testing there's no actual way of telling what they've caught unless it's severe enough to land them in hospital. And of course despite all the doom and gloom stories some people get over Covid-19 just fine at home.

Death figures, meanwhile, were initially VERY conservative, as only confirmed cases dying in hospital were included. So no test = not included, positive test but didn't die in hospital = not included. Then around April time the goalposts were suddenly shifted and anyone who had any possible hint of Covid-19 had it stated as their primary cause of death, regardless of anything else that happened. Both sets of data are amazingly flawed but ironically they probably balance each other out, the real figure being somewhere in the middle. And despite all that they're still more accurate than the infection rate data, which is basically "pin the tail on the donkey".

No state will “beat”New York, ultimately....both in raw numbers and in per capita.

It’s possible that some state might “beat” New Jersey.

Arizona and Louisiana will beat New York per capita probably within a week actually. The current numbers per capita are as follows:

New York: 2081
New Jersey: 1987
Louisiana: 1859
Arizona: 1849

Now remember that New York and New Jersey double in a year while Louisiana and
Arizona double in a month. So for the sake of estimate assume that Louisiana and Arizona are the only two states whose numbers will increase. It takes a month for them to double, so it will take a week for them to gain a quarter. A quarter of 1800 is 450 while the difference between their current number and New York is slightly less than 250. So yes they will likely beat New York in a week. Unless they start to plateau all of a sudden in which case I would give them two weeks.

As far as raw numbers, California, Texas and Florida all have more than 300 thousand cases while New York has just slightly above 400 cases. So since it takes a month for California Texas and Florida to double they will easily beat New York in terms of total numbers in a couple of weeks, probably less

Yo sum it up, I predict that in two weeks the following will happen:

a) As far as per kapita, New York will be behind Arizona and Louisiana but it will still be ahead of California, Texas and Florida

b) As far as total numbers New York will be behind California, Texas and Florida but it will still be ahead of Arizona and Louisiana

And, therefore, New York will no longer be number 1 no matter whether you look at total number or per kapita.

As far as *deaths* per kapita, New Jersey and New York will still be on top (interestingly enough New Jersey is ahead of New York as far as deaths go) but that is due yo the time delay between the illness and death.



But the difference between New York per kapita and New Jersey per kapita is only around 100 and 100/2000=0.05 so it's around 5%. Any state that can beat New Jersey can also beat New York.

Last edited by QFT on 16 Jul 2020, 10:13 pm, edited 1 time in total.

In raw numbers, no one will “beat” New York.

Some state might overtake New Jersey in raw numbers.

Per capita, you’re right, of course, when it comes to cases. When it comes to deaths, New York and Jersey still are way ahead.

But the sheer numbers for NY and Jersey compared to the rest of the states, until recently, was absolutely monstrous.

And NYC is off the charts as far as cities are concerned: 220,000 cases and 22,000 deaths—about 5 times in raw numbers 2nd place Chicago. Queens County alone has almost double the deaths of Chicago. So does Kings County (Brooklyn). The Bronx has a very similar amount of deaths to Chicago—slightly more, though.

California, Florida and Texas will. In terms of row numbers you have

New York: 404,775
California: 362,974
Florida: 315,775
Texas: 305,854
New Jersey: 176,501

Florida doubles every 2 weeks while California and Texas double every month. So in two weeks Florida will beat New York in terms of numbers for sure, and California and Texas probably will as well -- unless they suddenly plateau which doesn't look likely.



California, Texas and Florida already did.



Thats not what you said in the previous message. In the previous message you said



I am glad I was able to change your mind RE: per capita.



Yes they are. Thats because there is a huge delay between illness and death. So since New York and New Jersey *used to be* way ahead in terms of the illness, its no surprise they are *still* way ahead in terms of deaths.



"Until recently" is the key. Thats what makes it so interesting.



On that site that I mentioned in the OP, there is a map where you can scroll over with your mouse and it will tell you cases per capita for any given county. As it turns out, some of the counties -- in states other than New York -- have beaten New York for quite some time (probably around two months ago if not even earlier). There are a lot of counties with 5000 per capita or even 7000 per capita (and there are a select few with 8000+ and 9000+ per capita), while the state of New York is 2081 per capita and Queens county is 2927 per capita.

So the fact that other counties have beaten New York is the old news. *But* the fact that some of the *states* are about to do so is the recent news.

New York State actually has about 430,000 cases.

“Per capita” doesn’t really matter if there is, say, 3 deaths in a rural county.

The impact of 7,000 deaths within 100 square miles is much greater than the impact of 3 deaths within 100 square miles. And Brooklyn has over 7,000 deaths within about 65 square miles.

(Though 1 death within any amount of square miles has impact to the families and friends of the dead person).

What matters most is population density.

Where were you looking? I got my number of 404,775 from https://www.cnn.com/interactive/2020/he ... and-cases/ Where did you get your number from?

But in any case, our numbers are rather similar, so even if I were to replace my number with yours, I will still end up arriving at the same conclusions.



The total numbers don't account for population density though. You are assuming that the state has more people because it is more densely populated. But it is also possible that the state has more people simply because it has bigger area -- even though the density of people is around the same.

I guess one argument in favor of saying that its density would be the fact that some states are comparable in size while their population is vastly different. For example, California has 39.51 million people while New Mexico has 2.097 million people, so California times more people than New Mexico yet the ratio of their areas is probably around 2 or 3, but definitely not 20. So that would imply that the issue here is density rather than area.

However, this difference in density appears to be inconsequential, as evident from the fact that the per capita numbers *stay* the similar over the months. What you are trying to say is that if -- at a given point in time -- per capita numbers are the same then -- later on -- the state that is more densely populated will end up having higher per capita numbers. Well, this doesn't seem to be the case, at least not when it comes to comparing California and New Mexico.

One explanation I can think of is though is this. When you look at any given state, people aren't evenly distributted. Instead, they are living in the cities with plenty of empty spaces between cities. So maybe the area taken up *by the cities* in California is, in fact, 20 times larger than the area taken up *by the cities* of New Mexico and, therefore, the population densities of those cities are comparable. The fact that the average density of New Mexico is smaller is basically translated into the fact that the cities in New Mexico are a lot further apart from each other -- but that doesn't amount for anything since the distance between cities is far greater than the distance by which infection travels, no matter what state you go to.

So to sum it up, what I said so far is that the density that is important is the density of any given city; on the other hand, the average density over state is *not* important. Now, the rationale for looking at the total numbers instead of per capita would be average density over state. Since average density over state is not important, that means that we have no good reason of looking at total numbers and we should be looking at per capita instead. And this is, in fact, confirmed by the data that the states with similar per capita data remain similar.

Be it as it may, what makes the density even less relevant is that we are told about the doubling rates for each state. Your reasons to care about density is to "theorize" how fast it will transmit based off of it. But why "theorize" if we are actually "told" what the transmission rate is? What we are "told" is that New York takes a year to double while the average state across the country takes a month. So that means that New York is a lot less dangerous than you would think.

I meant the density of a particular area.

Very urban areas suffered much more extremely in all areas from COVID than rural areas.

More densely populated areas within urban areas suffered more extremely than less densely populated areas within urban areas.