davidgmills
But only give him a 38% chance of getting enough votes to get the needed amount of pledged delegates.

https://projects.fivethirtyeight.com/2020-primary-forecast/
Ye shall know the phony left by the gates they keep.
Quote 0 0
Fire With Fire
Hey, David

I know this sounds like snark as we have extremely different views on science, but I mean it as a straight up question.  Do you regard what Silver and other statistics based predictors do as "science?"  Analytics has dramatically changed how baseball and basketball are played.  Is that "science?"

This of course begs the question of what is science?  And although I think that might be an interesting discussion as well, for now I am asking for your view on statistical analysis as a way of predicting the future.  I'll buy that meteorology is science.  How hard does the "data" being analyzed have to be for the effort to be deemed science?
Quote 0 0
davidgmills
That is a hell of a good question.   And a lot of people much smarter than I have wrestled with it.  But here is what I know.  Let's start with the scientific method first and then compare it to statistics.

The scientific method is supposed to be about a process that is repeatable and duplicable.  I suppose a recipe is a good example of the scientific method.  If you faithfully follow the recipe, using the exact quantities, and cooking at the precise temperature for the precise duration, then you should get the same result over and over. 

My father, who was a professor of biochemistry and who taught biochemistry to graduate and medical students at UTMB in Galveston taught me a very valuable lesson as a young green lawyer.  I might have had my license six months.  I had been asked to take on a heroin possession case and I got a lawyer friend who had about six more months experience than I, to help me and be lead counsel.  My friend had the idea to file a motion to inspect and test the heroin and the judge granted the motion.  So I asked Dad to test it for me.  And Dad asked me if I was sure I wanted Dad to do that.  Being clueless, I asked why not?  He said, if I test it and it is heroin there will be no reasonable doubt about it.  What you want me to do is to criticize the state's test if you are not absolutely certain that it is not heroin.

The point he made from a scientific perspective is that faithful reproduction of a test is the only way to falsify it or debunk unless you intend to do an even better test to debunk it, which is what dad would have done.  So what I always look for in claims of debunking is first, was the original test described in such a way that its methodology could be faithfully followed?  If not then it does not meet the test requirements of the scientific method.  But if it is described in a way that it could be faithfully repeated, did someone faithfully repeat the test, and if he did, did he get a different result?  If he got a different result, then he could claim that he was unable to repeat the results of the first test.  Which then would require additional tests to prove or disprove repeatability.  But obviously the other way to debunk a result is to do a much better test than the original and get a different result.  What I so often find is that claims of dubunking never attempt a faithful repeat of a test or never attempt a better test.  And that is just blatantly obvious when you look at what has been done by a "debunker."

So to me that is the scientific method.  Of course it uses math in measurements and time, etc.

But for a whole host of reasons the scientific method is not able to be used for lots of things.  For one thing it can cost lots of money.  For another, it can take lots of time to do a test.  And time is a huge problem when the test could take hundreds of years.

And in situations like those, we have gone to using statistics as a substitute.  Sometimes stats are a bad substitute and sometimes they can be a pretty good substitute.  The major problem with statistics is threefold: finding a random sample to test (because it has to be random) and then having a large enough sample to have predictable results and having a limited number of variables involved.  And this has turned out to be a huge problem. 

There is a very well known 2005 paper by John Ioannidis an MD who looked into the predictability of the efficacy of drugs.  The title of the paper is a huge hint.   Here is a brief summary of the paper on Wikipedia:

Quote:
Ioannidis's 2005 paper "Why Most Published Research Findings Are False"[9] has been the most downloaded technical paper from the journal PLoS Medicine and is considered foundational to the field of metascience.[17] Ioannidis wrote that "a research finding is less likely to be true when the studies conducted in a field are smaller; when effect sizes are smaller; when there is a greater number and lesser preselection of tested relationships; where there is greater flexibility in designs, definitions, outcomes, and analytical modes; when there is greater financial and other interest and prejudice; and when more teams are involved in a scientific field in chase of statistical significance."


What he essentially found, was that out of 50 of the most prescribed drugs only 2, if I recall, were as efficacious as statistics had predicted.  Many had bad side effects that far surpassed any efficacy.

And if that is any indication about how good we are at predicting things through statistics, the clear answer is not very good.  Think about how much a drug trial costs of even a 100 people.  And really you probably need about a 1,000 which is the number often used in polling.  It is cost prohibitive.  But even that number is problematic if you can't really get a random sample of people.  And how do you do that with only a local population to use as a sample? 

And then there is one other nightmare -- how many variables are there?  In a drug test you essentially have one variable.

Here is another example.  Take the population bomb theory.  It has essentially one variable.  How many children does a woman have to birth to sustain a population?  The answer is 2.1.  So for years we were able to predict that the population would increase at a consistent rate.  But then something happened in almost all areas of the developed world.  Women quit having 2.1 births.  And now we know that we will never hit anything close to 11 billion at the end of 2100 as predicted.  It now seems that in much of the world deaths are outpacing births right now.  Probably we will at most reach 9 billion by 2050 and start a decline.  And we could not get the right prediction out of one single variable.

There are many things that we try to predict that have a huge number of variables, the most obvious is climate change. So from the perspective of the scientific method, we can't prove what will happen.  We can't even reproduce the greenhouse effect because a greenhouse has a solid glass ceiling.  Heat does not transfer through a solid the way it transfers through a gas.  So we can't use the scientific method to prove the greenhouse gas effect because we can not build anything that closely resembles our atmosphere from sea level to a hundred miles up or more. 

So we turned to statistics to predict climate.  But the variables are just infinite.  We couldn't even get population growth right on a hundred year times scale with one variable.  How could we hope to do it with an infinite number? 

Yet I as a lawyer think our society has every right to demand that we follow the advice of our best scientists or follow the consensus of our best scientists.  And I think that climate denialists do not really get this concept.  I may very well disagree with the consensus of scientists, but from a legal and political point of view, I think governments are obligated to go with the opinions of the best scientists or the consensus of the best scientists.  The denialist point of view is that the burden is not on them to disprove AGW, which is true from a scientific point of view.  It just clashes with what society wants which is to take the recommendations our best scientists make.

I don't really know whether that answers your questions, but if not, I am willing to take another stab at things.

And the "what happened" answer to the heroin case, is that we got lucky and the state had a chain of custody problem, and offered our client probation which he gladly took.

I also bet our views on science aren't as different as you think.
Ye shall know the phony left by the gates they keep.
Quote 0 0
deepestblue

Sanders vs. Trump:  Another Round of Lesser of Two Evils

 

Unlike the 2016 election where if one wanted the world to end in a matter of weeks they’d vote for Herself, but just like the 2012 presidential election, it will be lesser of 2 evils in the unlikely event Sanders is allowed to be the democrat nominee.

 

Trump – cutter of social safety nets

vs.

Sanders – opener of U.S. borders

 

The issue of illegal immigration and the desire for open borders on the part of the globalists is taken very seriously by a significant portion of those who are sympathetic to the messages Trump and Sanders were giving during their 2016 campaigns.

 

Sanders’ strength in 2016 was he maintained a laser focus on “kitchen table” issues. 

 

Now, Sanders has caved in to the misguided progressive wing and to the fake news media which blamed Obama’s border security policies on Trump, from the cages on down.  His caving is a politically calculated move that goes against the instincts he displayed throughout his career prior to 2017.

 

A major weakness of his in ’16 and ’20 was his inability to articulate any semblance of a foreign policy agenda, other than saying during this cycle that he wants to redirect some military spending to climate change spending and that the MIC is afraid of him, whatever that means, and I highly doubt they are. 

 

I read the foreign policy positions on his website which have almost zero depth. Compared to Tulsi’s positions, it’s night and day; she outlines very specifically how she is anti war in multiple sections; Sanders, not so much at all.

 

Sanders foreign policy section:  https://berniesanders.com/issues/responsible-foreign-policy

 

Gabbard foreign policy section:  https://www.tulsigabbard.org/tulsi-gabbard-on-the-issues/foreign-policy

 

Quote 0 0
deepestblue
In New Hampshire, incumbent Trump obtained more Primary votes than the last two incumbent presidents COMBINED.  NH is a state that has voted for a Democrat in every presidential election but one since 1992 (they voted R in the 2000 election - by the way, gee thanks for that one lol - those 4 electoral votes would have kept Bush Jr out of the WH).

[FJHm4ur] 
Quote 0 0