Wednesday, February 5, 2020

Groundhog Day


The actions of a small rodent don't predict the weather, but it makes for a good story.

First, an update:  The results of the Republican Iowa Caucuses are in!  Donald Trump has narrowly beaten former Hawaii Rep. Tulsi Gabbard for the Republican nomination.  While many in Iowa praised Gabbard's conservative views on many issues, they felt that Trump should be given four more years to build his wall.   Tough luck, Tulsi!

* * *

Journalists like to use unimportant criteria to predict Presidential Races, the economy, the stock market or whatever.  For example, they will say something idiotic, like, "The winner of the Superbowl has accurately predicted who will win the White House 50% of the time in the last 20 years!"   Sounds impressive, until you realize that is the last four elections, and even then, it wasn't much of an indicator.

There are others - how the stock market does for the first five days of the year indicates how it will do for the year!  And this has been accurate since 1982 - at least 60% of the time!   You might as well practice VooDoo.  It is akin to saying the weather is controlled by a small rodent in an obscure small town in Pennsylvania.

Why do journalists engage in this sort of nonsense?  Two reasons.  First, most don't have a technical education - they never took math in college or probability and statistics.   So anything mathematical to them is indeed VooDoo or supernatural    The second reason is obvious - clickbait.

When you have no sort of technical education, the world seems like a scary and random kind of place.  As a result, people devolve into a cargo cult mentality.  They adopt superstitions in the place of science.   And yes, I've met otherwise rational people who think that these "tells" or indicators actually have meaning and value.  That groundhogs predict the weather, and how the stock market does for three random days of the year predict the overall behavior.

And these same folks believe in "momentum" when it comes to elections.   Journalists are batting around other specious "statistics" these days, that in the last so many elections, whoever wins Iowa or New Hampshire ends up winning the nomination.   That may have been true, but that doesn't mean anything.  There isn't some sort of mystical control mechanism that prevents anyone else from getting the nod.   It is just correlation being confused with causation.

The results of the Iowa caucuses - if they are even completed by now - are anything but clear.   In an attempt to make the process seem clearer, they have succeeded in making it only muddier.   At the present time, Bernie Sanders can claim he won the "popular vote" while Pete Buttigieg can claim he won the most delegates - a situation oddly reminiscent of the outcome of last year's election.  Meanwhile, the remaining combatants can all claim that combined, their votes outpace the two top finishers - if Warren threw her support behind Biden or vice-versa, they would have won handily.   With so many people in the race and such a fractured outcome, it is hard to call a "winner" here at the present time.  Garnering one-quarter of the vote is hardly a clear win for any of the candidates.

But the media will have none of that.   Going into New Hampshire, the candidates have "momentum" and these two tiny States will pick the eventual winner!   The votes of the rest of the country don't count.   How odd, coming from the same party that decries the electoral college as "unfair".

The point is - and I did have one - is that we shouldn't let junk statistics influence us.  Which team won the Superbowl only indicates which team won the Superbowl.   Presidential elections are determined by who wins the most number of electoral votes - not by superstition, "tells", or "momentum".

Similarly, economic forecasts based on "the first three days of the year" are utter nonsense.   Far better data can be found in manufacturers orders, sales, profit-and-loss statements, and other hard numbers, which tell you more about the economy than any made-up "test".

But getting back to clickbait - why do people eagerly want to see the results of these arbitrary "tests"?   People get up early to see "Punxsutawney Phil" or stay up late on caucus night to see who "won" and get pissed off when no one does.  Why do we click on headlines that announce a coming recession or a bull market, based on little more than tea leaves or other arbitrary criteria?

It is a dark side of human nature, I guess.  And marketing people know this and play to it.  It is no different than so many of the "sponsored content" links which have questions in their titles or incomplete information, or "You'll never believe...." kind of things.   We are curious, and curiosity killed the cat.

Whoever wins the Democratic nomination and goes on to lose in November to Trump remains to be seen.   No sense calling the game before even the first play!