fd

Photo: The Independent 

A real-time event risk assessment tool created by Georgia Tech shows the risk you run of catching COVID-19 if you’re considering gathering with a big group this Thanksgiving.

The assessment tool breaks down the risk county by county in every state across the U.S. The chance of infection is based on the number of people attending an event, an adjustable figure, and if at least one person present at the gathering is COVID-19 positive.

The following figures represent the risk of catching coronavirus at an event with 25 people in attendance. The risk of contracting the virus heightens as the number of people attending increases.

Grant County – 42% 

Chelan County – 45%

Douglas County – 41%

Adams County – 61%

Kittitas County – 7%

Okanogan County – 12%

For access to the George Tech Risk Assessment Event Planning Tool, click here.

(10) comments

darling

Poor turkeys. Semantics. A Virus is a Virus. Common sense is not so common.

@the real JohnQPublic

I'm going to a memorial service for Mr. Tom Cluckers, so I'm exempt.

Desert Dweller

I would like to see the mathematical formula for the determination of those percentages.

EphrataResident

Unfortunately it seems that they used a program called Matlab for their modeling, which is used by scientists/engineers/mathematicians so unless you have the software, a few weeks to reverse engineer it, and the proper training it is impossible to determine how they arrived at their estimates (cost is ~$150 for personal, $2,150 for commercial license).

It appears that it is the same model that they developed back in March, with it updated last on April 18, 2020. They are feeding the model data from as late as November 17, 2020 from the New York Times (questionable data). However it is unclear what data is specifically being used, the criteria for said data, and how it is being manipulated. The model itself hasn’t been updated in 7 months, which is a bit concerning given the inaccuracies experienced around the time of the last update.

They did provide one assumption they made- the assumed undiagnosed cases. They assume that there are a minimum of 5 times more undiagnosed cases than there are positive tests (5 times more cases exist than are reported, possibly up to 10 times depending on whether they determine that not enough testing is happening in the area). Therefore in Grant County I guess they estimate 18,690 actual cases (or 5 X 3,738) so far. It is entirely possible, if not highly probable, that a significant amount more people have had the virus asymptomatically than have been positively diagnosed. However, this actually works against their narrative. The known case fatality rate in Grant County right now is around 0.86% (~1 in 120 chance of dying). If their 5:1 assumption holds true, that means the estimated case fatality rate would be ~1 in 600 since the number of cases would be 5 times higher (deaths remain the same, regardless of asymptomatic case assumptions). That assumption would also mean around 20% of Grant County has, or has had, COVID. It is unknown if they are taking herd immunity into account, but it’s doubtful since their modeling was last updated before herd immunity would have played a major factor.

Intuitively, we can surmise that 42% is way higher than the actual risk (as proven by the inaccuracies of all models to date). It should also be noted that, per their own admission, “The model is simple, intentionally so, and provided some context for the rationale to halt large gatherings in early-mid March and newly relevant context for considering when and how to re-open.” In other words, their goals were, and still are, to justify shutting down gatherings. That’s not science- that’s called generating confirmation bias.

Sagacious Lu

It looks like you can arrive to a similar result just using a simple cumulative binomial probability. Any spreadsheet will yield the result - the key is the active infected rate. Some numbers:

A B

1 Data Description

2 1 Number of successes in trials

3 25 Number of independent trials

4 1.80% Probability of success on each trial

Formula Description (Result)

29.0997% Probability of exactly 6 of 10 trials being successful (0.205078)

900 active

1800 probable active

100000 total population

1.80% probability

Where I have defined "success" as expposure to COVID-19.

The Tutorial page from GA Tech gives more info.

https://covid19risk.biosci.gatech.edu/

Now, of course, the binomial distribution assumes independence of trials, so, given that these are family gatherings,that assumption may be inappropriate.

EphrataResident

According to their website, they assume that the actual case number is 5 to 10 times higher than the tested rate... so probability for independent "success" would be 4,500 if that held true, right?

It should also be noted that symptomatic members of households would theoretically stay home, along with all other members in their household. A sort of canary in the COVID mine, as it were (which is where I think your statement about independent trials applies).

At either rate, you can manipulate any set of data with statistics to have it show what you want it to. In this case, their stated goal is to justify lockdowns, not accurately forecast infection rates.

Desert Dweller

That was the point I was making. It all comes down to who decides what information to input into the formula, and how you weight the data.

Well put by the way Ephrata.

EducatedAmerican

Thanks for you very detailed and EASY to understand post. I wish I would have had you as a professor at WSU for my statistics class.

Angry old man

smh stay home and tv dinners

darling

Good idea

Welcome to the discussion.

Keep it Clean. Please avoid obscene, vulgar, lewd, racist or sexually-oriented language.
PLEASE TURN OFF YOUR CAPS LOCK.
Don't Threaten. Threats of harming another person will not be tolerated.
Be Truthful. Don't knowingly lie about anyone or anything.
Be Nice. No racism, sexism or any sort of -ism that is degrading to another person.
Be Proactive. Use the 'Report' link on each comment to let us know of abusive posts.
Share with Us. We'd love to hear eyewitness accounts, the history behind an article.