My “count the doublings” predictions didn’t account for some things that have since come out.
The number of doublings until peaking is less. Which is good.
The number of asymptomatic people walking around is much higher. That was bad when social distancing, etc. was not in force everywhere. But it will be good down the line.
It’s getting clearer and clearer the US will relax restrictions far too soon so the “plateau” will be higher and last longer. That’s bad. Really bad.
So maybe 200k-500k now would be my guess for this year unless things go nuts again in the fall. Then whoosh.
It could only go that way if we relax social distancing prematurely and have a rebound epidemic.
I predicted 900k dead in 2020. I’m thrilled to say it looks like I’ll be very wrong about this unless our political leadership leads us into a rebound epidemic by being overeager to relax isolation in time for the election. I really hope to be wrong about that, but I’m not seeing any voices of reason to mitigate the malign influences.
I am getting increasingly optimistic. My guess at 32 000 around April 23rd I’ll keep but now I hope the rate will continue to drop and we won’t get to 40 000 by the end of April. My fears of southern/western metro areas getting really bad doesn’t seem to be happening.
I’m sticking with my original prediction of less than 100K by the end of the year, and possibly considerably less. Whatever happens later in the year, we will be much more ready for it in terms of testing, equipment, facilities, treatments, etc.
The numbers one needs to predict things are still fairly shaky.
E.g., for the week ending April 3rd, England and Wales had 3,475 dead with CV cited in the cause of death. Yet there were an excess of 6000 deaths that week over what would be expected. The vast majority of those have to be ascribed to CV directly and almost all indirectly.
Furthermore, the number of flu deaths could be lower than usual thanks to all the measures taken. So the CV toll would be higher still.
So around a factor of two in underreported deaths. And in a place that’s doing a better job than the US on tracking, testing, etc.
My insightful intestines advise me to stick with my prior [del]estimate[/del] WAG of a close order of 2 million excess deaths in USA 2020. Since not all stiffs will be tested, we can’t really know how many are caused by COVID and how many are merely collateral damage. Did they die FROM the virus, or WITH it, or were they down the triage scale?
So those thoughts on record I’d like to request some predictions of what those numbers by the end of the year would mean in a nearer term, such that the predictions would be in a nearer term be subject to falsification.
Feel free to use end of the month number, or date when 100K will be reached, or both.
CarnalK already went and I’ll mostly concur with about 36K by the 23rd, 40 to 45K by the end of the month and not hitting 100K in the calendar year. I hope to again have erred to the high side. (And would take under on an over/under bet.)
Please note, decent excess mortality stats won’t be available for a long time so these predictions are based on, have to be based on, deaths attributed to it in real time.
It looks increasingly likely that my guess of a half million is going to be too high. But I’m going to refrain from any revisions. I don’t want to later say “well, my third guess was better than my first, second, or fourth, but not as good as my fifth…”. Let it ride and see how it goes.
Best to not overanalyze a single day’s data, one direction or the other. Definitions changed today so new “probable death” cases were thrown in I think.
I found this thread awhile ago and I’ve been holding back. Sometimes I’m encouraged by what I’m seeing and other times not.
I’m going to go with 100K-200K. If my math is right it took 47 days from death #1 to today, about 34K. We have 200+ days to the end of the year. I know it can be a geometric progression—but people may get spooked and back off for awhile and flatten the curve. But IMO it doesn’t last…there are economic realities to consider.
Worldometer says of the resolved cases, where people either expired or recovered, 21% have died. We’re approaching 700K cases in the US, which would translate to about 150K dead. But OK we’ve got GM gearing up to make ventilators now and hopefully we’re buying some time?
I don’t think this is a useful way to figure out what percent of people who get this disease will die. Most recovered patients aren’t going to show up in the “recovered” column on Worldometer because, most of the time, nobody is really keeping track. If you spent time in a hospital, you might get listed as “recovered” after you’re discharged, but if you had a mild case and got sent home to self-isolate until your symptoms disappear, nobody is going to follow up in most places.
The WHO thinks the fatality rate is 3.4%. Other scientists think this is an overestimate and it’s lower, but I don’t know of any credible authority that thinks it’s higher outside of certain very specific circumstances, like healthcare systems being overwhelmed.
This is the difference between countries which managed to keep it contained and the ones which were forced into mitigation. Taiwan doesn’t discharge patients who have minor symptoms or even those who are asymptomatic. They are kept in isolation in a facility.
Of course, once a country gets overwhelmed then it may no longer be possible to do this unless many, many more beds were built.
I will grant you, a lot of people are probably not diagnosed and so they recover. We can’t know what those numbers are…by definition, they’re not diagnosed. Maybe they have very mild symptoms that they easily shrug off or maybe they self isolate for a day or two and they’re better but they don’t go to the hospital.
The number of people who DO get diagnosed may be because their symptoms are so severe that they must seek attention. If you’re in THAT group, the survivability is much tougher. Well in the US we have a big number of THOSE.
If I work the problem backward, how about this? 100,000 dead @3.4% would mean about 3 million infected. Right now we can account for about 700,000 confirmed. Let’s say an equal number exist that are NOT confirmed and we’re about half way to 3 million total.
I hope we see fizzling out in NY and other hot spots and I hope we don’t get new hotspots elsewhere. I also hope the second wave isn’t bad and that we don’t prematurely return to normal life. But then these protests in places like Michigan make me think this process is going to have some severe hiccups.
12/31/20 = the OP’s choice…“by end of 2020.” That’s about 250 days.
Predictions of 45k by the end of April look well off, its 12 days to go and there are 39k.
I do not know how useful it is to guesstimate based upon deaths per million population but currently US has rates about 25% of nations such as Spain, Italy, Belgium, and when you look at others who still are on the increase such as UK Netherlands they seem to be following similar death rate curve.
That might suggest for the US that by the end of the month it might be around 80-100k in the US by May1st and then its a case of how quickly the death rate slows once the peak has been reached.
In this wave perhaps 160k but that is only if things do not suddenly increase, and those protests and church gatherings haven’t yet fed into the infection rates - if worst happens and those events do cause more widespread infection then maybe 250k.
As for total for the rest of the year, even more unpredictable and probably depends upon political decisions such as when the election comes up. Over optimism and unlocking too soon could lead to much higher numbers, it only takes one or two states to do this and all the work in the more prudent states will be undone, it really takes every state to work together and I’m not sure the US political system is capable of doing that - it doesn’t seem to be fit for purpose for having a uniform coordinated national response, especially given the national leadership which has completely different agenda.
Hard to know how far off the predictions are since they changed the accounting methods and did a sudden one day add of 6200. A couple bad days after that but deaths per day isn’t really going up, istm.
No question that the hypothesis had not foreseen or included the change in case definition that shifted the curve up (a similar event happened dramatically with China’s numbers early on), so yes the prediction the hypothesis makes will likely be off by the amount the definition change adds in over the two to three days it takes to get those now counted case numbers in. And those numbers really should be counted.
I accept the growth curves prediction will be off by somewhere around the amounts the expanded case definition add in. If it ends up that range though the hypothesis remains not yet falsified.
casdave, would accept your read of the numbers to be falsified if the end of the month numbers are significantly below your lower bound?
But had people behaved as usual with no distancing it might have happened that way.
For the two weeks ended yesterday, April 18, the average daily growth in number of cases was 6.4% (doubling about every 12-13 days) and the average daily increase in deaths was 11.5% per day. (doubling about every 6 days)
At 6.4% daily growth in cases we’d reach 1 million US cases with the next week and at 11.5% increasing death rate we’d reach 100,000 deaths by th 27th. However, recent percentage increases have been 4.5% for cases and 6% for deaths. If those rates keep decreasing we may stop at a bit over a million cases and under 75,000 deaths.