Friday, October 2, 2020

Unknown unknowns: All our crystal balls are cracked

 

     My wife and I read to each other almost every day. We take turns. I usually read something relatively light to her -- some fantasy or science fiction, perhaps a mystery. Events around me are too serious and I need some escape from them. On the other hand, she often chooses books that have much more serious themes (this is also true for movies that she chooses) and occasionally non-fiction. (Though we have just run across a quote that claims that fiction often contains more truth than non-fiction. That may be true, especially about interpersonal relationships.)

     At the moment, it is about "black swans". The book, The Black Swan: Second Edition: The Impact of the Highly Improbable, by Nassim Nicholas Taleb is a discussion of those things that really affect the final results. It is related to "Chaos Theory" but not really quite the same thing. The famous quote about Chaos Theory is that a butterfly flapping its wings on the other side of the world may determine whether or not a thunderstorm near you will happen.

     That of the "highly improbable" is more focused on the "unknown unknowns". A lot of the time, we try to bring into account -- plan for emergencies -- about unknown events in the future. Earthquakes, tornadoes, fire in the building, a key employee being hit by a bus on the way to work -- these are all known unknowns. We don't know that they will happen but we can envision them happening and attempt to make plans in case they do happen.

     Life is infinite and unpredictable. It is impossible to envision all the things that could happen. Even though a pandemic has been discussed as inevitable for at least the past 30 years (maybe earlier and I just was unaware), no one expected it to hit the world in 2020. It doesn't always have to be such a momentous occasion. I go to the barber shop and a siren blares right outside the door just as the barber is making a trim of my neck with shears. Ooops. All of a sudden, I have half of a Mohawk.

     I consider the stock market and the weather both in the category of being made up of a lot of unknown unknowns. Sure, for an individual company, there may be known factors that would/should influence the price. But how often do we read column after column of articles about why something happened the previous day -- and no articles the day before yesterday about those same factors and how they will affect the market that day? I appreciate the extensive models, training, and hard work of meteorologists but they are still trapped in the world of percentages. The following day, they can give you all kinds of reasons why the previous day's weather occurred.

     Most humans do not like to not know what is going to happen. They greatly desire the illusion of control. So we have a lot of prognosticators -- people to guess the future. And, they are all in the same situation as the weather forecaster (whether they realize it or not) -- they can do all kinds of things, use all kinds of tools, check all kinds of data -- and they can only give percentages. The closer in time, the better the odds. Part of the purpose of the Agile methodology, which was discussed in an earlier blog, is to keep allowing the focus, and priorities, to keep changing as new information comes in. It reduces the time to event and, therefore, increases the percentages. But those percentages never reach 100%.

     In my previous company, we underwent a periodic future product analysis to determine our priorities in development over the coming year. Part of that was to determine a theoretical Return On Investment (ROI). Perhaps some larger companies, with greater resources, achieve a better percentage of accuracy. Ours was not that good.

      If they ever say 100% you are entitled to have great doubts. Like a person tossing a coin into the air and guessing heads or tails -- and it lands on its edge or a bird grabs it mid-air -- there is always something (no matter how unlikely it may seem) that can arise.

Interrupt Driven: Design and Alternatives

       It should not be surprising that there are many aspects of computer architecture which mirror how humans think and behave. Humans des...