Monday, March 23, 2020

GIGO.

Ah, now, some interesting facts about Italy's pandemic begin to emerge. As Russians say--it doesn't matter how many, it matters who (pun is unintended) counts. As Ria reports (in Russian) there is a teeny-weeny "issue" with the high mortality rate (emphasis on "rate") in Italy because, as it is worded, of the "method of testing" (in Russian). Namely, the issue I am writing about non-stop and appeal to everyone to wait for the larger data set. Well, guess what: as deputy director of WHO stated today, Italy changed the method of testing, due to, get this, shortage of test kits and counts only "confirmed" cases against deaths and here is the direct quote of this WHO deputy director: if the mortality rate would have been calculated against the number of actually infected, which, and I quote, would raise the number of infected many times--the mortality rate would have been much lower. Pardon my French, no shit! 

As was predicted (well, by me) we are dealing with a classic GIGO--Garbage In, Garbage Out. It absolutely doesn't mean that serious measures against Covid-19 shouldn't be taken, they should. But as it is always the case in operational assessments in military, one thing you always need is a reliable verifiable data. One cannot base decision on unconfirmed and un-weighted, that is processed for their probability of being accurate, reports based on opinions of limited number of observers (sensors), military people invented Sensor-Data Fusion for that and have well-agreed upon mathematical apparatus which accounts for probabilistic assessment. I can report that my forces destroyed 50% of enemy's forces based on observations of the front line, but they will tell me nothing (unless I am aware by means of ISR) about the state of reserves which are moving and, once reserves are counted, by percentages describing super-pooper "victory" will drop dramatically from 50% to much lower number.  Even in the old times of steam when navigating oceans, when using Sextant to get a position line form individual celestial body, navigators would take usually five-six measurements within equal short intervals of time to average (that is partially resolve uncertainty) the measurement before applying it. Similar algorithms today are realized in optronic means of navigation and targeting--non-stop measuring, with constant averaging. Of course, today it is done by modern processors. 

This all doesn't mean that Italy is not in danger, but it does raise the issue which I raised few days ago: we are yet to learn (hopefully) how WHO arrived to a decision to declare a pandemic globally. There ARE questions about this decision and the panic which ensued. Russia, meanwhile, is operating the air bridge to Italy, the tenth military cargo landed just very recently near Rome, and efforts to contain and prevent will continue in earnest.  But after all is done and over with, we must ask ourselves a question even maroons in NYT are beginning to ask themselves--didn't we all go a "bit" too far? Meanwhile in good ol' USA Congress, as is expected, cannot do anything of value, being embroiled in.... normal political process. 
Isn't it wonderful? I guess that is how it is supposed to work. Right? It is finally raining today around Seattle, after two very dry weeks. One of the main ways Covid-19, or any other virus, spreads, apart from direct contact, is a an air-droplet way. Rain is a welcome relief from early Spring allergies and raising humidity to levels which help blocking all kind of airborne shit. I am in the mood for the melody.

No comments:

Post a Comment