Monday, February 01, 2016

Math questions that come out of the Iowa Caucus

Link for info I refer to here

Jim Gilmore, republican, has 0% of the vote. Does that mean that literally NOBODY voted for him?

Hillary beat Bernie , but BOTH get 21 delegates. I hardly call that a win. I call that a tie.

Why do we refer to Hillary and Bernie by their first names, but most of the republicans by their last name. The only exception is Jeb! who we call Jeb to dist from his brother. Also, I think he wants to play down his relation to his brother. Not that it will matter.

Cruz/Trump/Rubio will get 8,7,6 delegates.(This might change but not by a lot). I'd call that a tie. Later states will be winner-take-all which make no sense in a race with this many people (though it may go down soon-- Huckabee has already suspended his campaign which seems like an odd way of saying I quit). But in winner-take-all states there will be real winners.  Why are there winner-take-all states? It was a deal made so that those states wouldn't move their primaries up.

This is a terrible way to pick a president. I don't mean democracy which is fine, I mean the confusing combination of Caucus's and Primaries, with some states winner-take-all, some by proportion, and Iowa and NH having... more power than they should.  This was NOT a planned system it just evolved that way. But its hard to change.

If you are a registered Republican  but want Hillary to win then do you (1) vote for the republican you like the best, or (2) vote for the Republican that Hillary can most easily beat. The problem with (2) is that you could end up with President Trump.

Stat analysis of polls and looking at past trends have their limits for two reasons:

1) The amount of data is small. The modern primary system has only been in place since 1972.  Some nominations are incumbents which are very different from a free-for-all. The only times both parties had free-for-alls were 1988, 2000, 2008, and 2016.

2) Whatever trends you do find, even if they are long term (e.g.,  the tallest candidate wins) might just change. The old Machine Learning warning: Trends hold until they don't.

Most sciences get BETTER over time. Polling is a mixed bag. On the one hand, using modern technology you can poll more people. On the other hand, people have so many diff ways to contact them that its hard to know what to do. For example, its no longer the case that everyone has a landline.

Is this headline a satire?:here

AI project- write a program that tells political satire from political fact. Might be hard.

My wife pointed out that

Hillary WINS since she didn't lose!

Bernie WINS since an insurgent who TIES the favorite is a win

Cruz WINS since... well, he actually DID win

Trump WINS since his support is mostly real and he didn't collapse. And he was surprisingly gracious in his concession speech. (This is the weakest `he WINS' argument on this list)

Rubio might be the BIG WINNER since he did way better than expected. Thats a rather odd criteria and it makes people want to set their expectations low.

SO--- like a little league game where they all tried hard, THE'RE ALL WINNERS!

 The American Public--- not so much.


  1. A (very) minor point: Huckabee suspending his campaign instead of quitting is standard procedure. Candidates do this for legal reasons (campaign finance laws).

  2. why did Hillary get 22 delegates and Bernie get only 21 if they tied within .2 percentage points and there are 44 delegates available?

  3. Rubio went up pretty fast and then collapsed pretty fast. It is going to be very hard for him to recover from this.

    Hillary is bleeding and we might have another 2008 by May.