New model says 200k dead by Labor day tbqh
That sounds reasonable?
And with the fall/winter wave. I could see half a million bodybags. Note that i had called exactly that months ago
It would be a shame to see another huge outbreak in NYC.
110M million to 1.8M.
If you were worried about 1.8M infections, how would 110M have made you feel?
It's always good to scare the out of people who don't have any common ing sense.
Thoughts and prayers
That's simply not true. Plenty pick the model that best supports their narratives. So if you think the infections will be 110M, and they are 100x fewer, what good is a model? Couldn't you just ask a 3rd grader to think of the biggest number they could and go with that and probably be more accurate?
Or at the next trump rally tbqh
Nobody made a claim the data was wrong. That's you premise.
Of course it's true. We're arguing from the point of parties being interested in mitigating the problem, not people trying to advance narratives. You don't need a model for lying, you can just make up any number.
Your complain made sense if you disputed it when it came about, not months later. Under what conditions was that 110M estimate reached? Let's see the parameters and the data when that model was put together, then we can discuss margins of error or sampling errors.
What model do you have a beef with right now? Pick a model right now, and call it bull like you're calling the 110M right now with the current data set, then we can evaluate.
Everyone is trying to advance a narrative, both sides. You don't need to make anything up. You just pick the model with the numbers you need. If you want to scare people, you say 100million. Then when it's 1.8 million you look like a hero. If you want to calm people, you show the cherry picked model that supports your narrative. The left wants to show Trump has killed every human on the planet and the right wants to show Trump rode in on a white horse (with a black name, let's be fair) and saved the world from the evil Chinese.
I don't have a beef with any model. People calling for concrete actions based on faulty modeling and insisting the models are correct, those I have a problem with. When you're right - blame it on Trump. When you're wrong - blame it on the modeling. When is it ever just you? (rhetorically speaking)
Models that are wrong by a factor of 100x are useless, and even more they are likely dangerous.
(anecdote) - I used to do business with some South American countries. Each quarter they'd have to project their earnings and each quarter they'd blow away their projections by as much as 200%. Now that seems great, they are out performing, but what's really happening is they are sand bagging the predictions to beat them easily, to both remain employed and to raise their own stock prices. Their predictions were more than worthless, they were costly. I didn't have an issue with the models. I had an issue with the people making decisions based on what was obviously highly questionable predictions, because it benefited them personally to accept them.
Last edited by DMC; 06-02-2020 at 08:48 PM.
Chanting "I can't breathe" may literally turn into their last words
Picking models for narratives is a double edged sword, and I think we can agree with that. It actually happened to Trump before he eventually had to recognize that 100k was well within the ballpark, and this was serious. That's why nobody uses models from twitter.
There are true and tested models for infectious diseases that have been used for years on Ebola, Flu, SARS, etc... and they're only as good as the data they're fed, AND all of them include a margin of error.
So if you were doing nothing, they will report 100M, if you tell it you've been doing social distancing, then it's going to come down a lot. It's not magical thinking, it's called mathematics. When you have a virus with an exponential growth profile, the numbers escalate pretty quickly, and what you do to mitigate that, or if you have a vaccine, or if you have a therapeutic treatment, or if the virus is affected by weather, etc they all influence the numbers over time.
It's incredibly easy to dispute statistical models in retrospect if you don't know how they were calculated to begin with. Thus my question on who has a beef with which model right now, because as soon as a couple of weeks have gone by, weather changes, these protest episodes happen, and then you have to feed that data in and get new numbers.
Was it wrong two weeks ago? No, it was within the parameters and information it had at the time. Does the concrete decisions made on those values 2 weeks ago were somehow incorrect because now the model looks different? No, they're not, they were informed decisions based on the information you had available at the time.
We also make decisions that we know will influence the model in positive ways, like social distancing, mask usage, etc, and then we wait for both reality and the model to reflect that.
But wasn't Trump supposed to know and thwart it all before it took hold?
There's more than one model. That's my point. I can use the more extreme example or the most conservative estimate depending on what echo chamber I belong to.
She's white, ing Karen. Of course she'd say that. Forced herd immunity imo.
They're all in the same ballpark with the same data though. None of them are 15 down to zero.
And yes, what you did in the early phase of contagion also has an impact on the overall longevity, especially when contagion is exponential. If you let pockets spread freely early on, you're probably going to be battling with this longer.
I don't expect Nancy the hairdresser in Yuma, AZ to understand how models work, she'll probably get the initial estimate, and then complain two months later how the model hit or missed. But I know you know better.
I'm not even arguing that some outlets might A) weight the error rate towards more dramatic/conservative or B) outright lie (at which point arguing about the 'model' is a waste of time).
I'm discussing making decisions with sound model/data, which I believe this administration did in large part thanks to Birx scaring the out of Trump by telling him he could have 2 million dead americans in his hands. Was she wrong? I don't know that she was when that was said and we were doing nothing but downplaying it. I certainly would like to see the data/estimates back then.
Not sure what ballpark you play in. 110M vs 1.8M seems like maybe you're not playing baseball. Are there 18 bases?
They don't have the same data. They have the data that someone plugs into the model, cherry picked data from cherry picking people, each with their own agendas.
I'll even advance it could very well be the same model. You're simply comparing two different points in time with different cir stances. I suspect one is without doing absolutely nothing, and the other is with lockdowns, distancing, masks, etc.
Projection of data for one and the other are very different, because cir stances and mitigation are different. But we didn't do nothing, we actually did all the other stuff, and the model actually held up pretty well.
I'll have to look back at one of my posts in April when I said, looking at the reference model being used by the government, that it looked like we would be reopening around June 1st. That was actually fairly spot on.
We can all agree Trump's model was way off.
Fergusons death predictions have happened for years. The model has been highly criticized.
Imperial are trying to have their cake and eat it. Reports of random results are dismissed with responses like “that’s not a problem, just run it a lot of times and take the average”, but at the same time, they’re fixing such bugs when they find them. They know their code can’t withstand scrutiny, so they hid it until professionals had a chance to fix it, but the damage from over a decade of amateur hobby programming is so extensive that even Microsoft were unable to make it run right
But this isn't the point, the problem of current manipulation of data only serves to extend fear and mislead the public with daily cases and daily fatality numbers. Not really much you can argue with if you believe in science or math
Then bring a better model. It's easy to criticize and not provide a better alternative. Microsoft worked with the University of Washington, which uses a similar model, on the visualization tool that ended up being this site:
https://covid19.healthdata.org/united-states-of-america
which IIRC, is the model the US government ended up using. And it's been pretty spot on. We discussed some of the numbers where off for hospitalization at some point, and eventually, as more data came in, it was adjusted. That's just how statistical models work.
This is a conspiracy theory though. Because I can ask you to back up the contention that there's manipulation of data, but you clearly cannot.
Make up your mind man. Before these protests you were all about life having to go on and just open everything up.
Unless Vy thought that we would do absolutely nothing, you're basically saying he took the most liberal estimate and ran with it. How is that any different than what I said? It makes you wrong for two reasons (not you specifically), 1) wrong in numbers 2) wrong for the negative nancy outlook on American response
There are currently 5 users browsing this thread. (0 members and 5 guests)