A Breakthrough in Modeling: Informative Censoring

The Chloroquine Wars Part XCV

"Random numbers should be generated with a method chosen at random." -Donald Knuth

Nobody could possibly realize how smart Donald Knuth is until they've tackled a problem whose answer is not simply statistical, but actively hidden by other human beings with the aid of automated computation.

Working on a challenging modeling problem with the help of a friend.

Thanks to Mark Reeder, I think I have the final piece of a tremendous puzzle. I believe I can explain essentially all aspects of the conflicting data over vaccine efficacy, and in a way that will be convincing to good, honest, and open data minds.

On a practical level, I will explain how informative censoring---a concept that seems relatively recently named---is one of the most important lessons for a financial trader to understand. I'm often asked for lessons in how to trade, and I usually give specific answers about past scenarios, some real explanations about fundamental statistical methods, and explanations of structural weaknesses in financial systems, but I know I'm always short-changing the listener and I'm up front about that. But ultimately, the concept of informative censoring relates directly to the processes of identifying some of the best trades.

I may put together a spreadsheet and then a video explaining the modeling process, but frustratingly that won't happen quickly because I have jury duty this week and so few people are showing up that there isn't even room to dismiss even somebody like me with N-sigma unusual views.

To be honest, I would probably enjoy sitting on a jury, doing my best to help society control interpersonal conflict and harm. Most of the time. I might one day be one of those old men who strolls over to the courthouse and volunteers for grand juries once every few moon cycles, and writes about the process. A good court system, which we have [in pieces] still, is part of the framework of the kind of societal machinery that promotes better lives lived, on the whole.

But this week, the task feels like a waste of potential value of time. Sometimes random is just random.

If you want to be follow what's coming in future articles and maybe a video soon, read the paper linked to Mark's name.