Thursday, September 15, 2022

Dust bowls, passing of an era with Elizabeth II, watchtowers, permutations and other things dreamt of

On local oldies radio heard Jimi Hendrix playing his masterpiece, All Along the Watchtower, the other day. Dylan wrote the song 1967 or so, and Hendrix recorded it 1968. I recall at the time being mesmerized by completely unique electric guitar riffs and audio processing, Hendrix more or less creating the acid rock style that was soon taken up by other players for decades. The lyrics were another nod by Dylan to the Old Testament apparently, he probably being familiar with those verses (like many of us of past years, the words of the Bible are an eloquent underpinning to much of our thought, form and substance). Isaiah Chapter 21, verses 5–9 (KJV):

21:5-9 Prepare the table, watch in the watchtower, eat, drink: arise, ye princes... For thus hath the LORD said unto me, Go, set a watchman, let him declare what he seeth. And he saw a chariot with a couple of horsemen... and he hearkened diligently with much heed: And he cried, A lion: My lord, I stand continually upon the watchtower in the daytime, and I am set in my ward whole nights: And, behold, here cometh a chariot of men, with a couple of horsemen

I actually had tears in my eyes when I heard of Elizabeth II's passing and watched some of the many retrospectives on her. For me, aging and despairing at the course of humankind, it was more than simply the death of a good-hearted queen, monarch of a country which I still feel a great affinity for (as many of the original colonists and their kin). It was yet another loss of a type of nobility and integrity that is vanishing now. I mourn the death of both. It was telling to hear American news folks somewhat puzzled at the quiet reverence of thousands of Brits who gathered to see her off---the type of people who left Britain for the New World were definitely not the type who remained!

I watched a PBS special on the 1930's Dust Bowl climate disaster recently. I was fascinated with the reaction of the inhabitants of the Oklahoma/Texas areas affected, the denial and attempts to persist in their previous habits and life: 

This is of course highly relevant as I watch a global catastrophe occuring now, almost weekly damage and disruption of America from wildfires and vicious weather (well said by California governor Newsom to be "nature's fury"), flooding, destructive winds, rising sea levels inundating coastal areas, roasting humanity with unheard-of heat. A core of Americans react this time with the same denial, but with the added viciousness of the child of poor character encountering limits. And still we have the counterbalance of the perverse academia-media axis who "Who feel that life is but a joke" (from All Along the Watchtower).  Four horsemen are approaching, and the wind indeed has begun to howl!

I have spent recent weeks studying combinatorics, the related topic of permutation testing in statistics being as chaotic and imprecise as mud-wrassling a greased-hyper-pig, as we used to say in my hippie-dog days late 60's southwest Texas (well, as I would have said if I had been devoting much time to the problem then). Permutation is the process of producing new orderings of a sequence, think of quickly switching the order of three cups under which one hides a treat while trying to test your dog's intelligence (if that is still permitted, after all, the dog has its own cultural traditions). 

I ran a simulation of the random reordering of the differences in height of Zea mays plants grown by Darwin to test the superiority of cross-bred over self-fertilized (I have mixed feelings about the relative merits of both at this point in history, but that is another matter). This was suggested by Ronald Fisher in his 1935 Design of Experiments, thus initiating the subject of permutation testing, though, to my continuing irritation, subsequent practitioners tossed precise usage of mathematical terms out the window and loosely referred to any shuffling of data and labels as "permutations." 

The idea is that if you have a true experimental effect, in this context that a cross-bred plant will grow a bit higher (and attain a higher standard of living?) than its poor self-fertilized competitor, if you randomly change the signs (plus or minus) on the measured differences in height and count how many times the sum of the differences (one way to do it; there are other test statistics possible) is as extreme as that originally observed, then you can estimate the probability that you are seeing a random effect rather than a true experimental result. On the following plot I mark the observed original statistic with a vertical red line. The blue histogram plots the count frequency of randomly shuffled sums (sums of randomly alternated signed height differences), most being much lower, suggesting that our observation is unlikely to occur by chance:

To do the analysis and create the data used in my plot above, I pulled (extracted the computer code of interest and modified it to accomplish my limited goal apart from the original package) some Python code from a software package by Kellie Ottoboni (Ph.D. from Berkeley) and colleagues. The yellow curve plotted over the top of the blue histogram is a plot of a normal distribution with the mean and standard deviation of the histogram data, for comparison. By the Central Limit Theorem, if you take enough samples of a population, it eventually begins to look like the bell curve plotted, i.e., it converges to a normal distribution. Randomization techniques like "permutation testing" handle the case when the samples tested may not be normally distributed, i.e., they allow you to make inferences about the data without making any assumptions of normality (probably not a good idea to assume anything is normal these days, but that is another matter also).

I mention Kellie in particular because a paper she wrote with two collaborators, An Empirical Comparison of Parametric and Permutation Tests for Regression Analysis of Randomized Experiments, was very useful to me in bringing me up to speed on this subject. Coincidentally, one of her Ph.D. advisors (who also worked on the Python code cited earlier) at UC Berkeley, Philip Stark, had written and posted one of the first online courses taught at UC Berkeley, Introductory Statistics, which I had studied in February 2022. Stark writes well and has a good sense of humor, e.g., the epigraph for his online course is "statistics means never having to say you're certain."

I cringe everytime I hear a new announcement on public radio attributed to "new research" or "science" because it is inevitably some dubious or trivial result reflecting either an outright distortion or misuse of a very small statistical significance. In Practical Regression and Anova using R, Julian Faraway quoted St. Augustine (4th century):

The good Christian should beware of mathematicians and all those who make empty prophecies. The danger already exists that mathematicians have made a covenant with the devil to darken the spirit and confine man in the bonds of Hell. 
In Confessions and City of God (two major works by Augustine) I could not find this purported quotation, although it is more or less a good summary, in more metaphorical phrasing, of the many discussions Augustine did offer on the absurdity of astrology, which at the time the practitioners of which were labelled "mathematicians." 

Speaking of covenants with the devil, I feel sorry for Liz Cheney, who must now realize that all those years she believed she and her father naturally obtained roles as leaders because of their integrity and intelligence were years of delusion, illusion. In fact the population which had elected her (and now worships the antichrist and continuing polluter of the American Republic) is largely bestial and moves this way and that with no more higher thought than that of cattle in a herd. It is ironic, of course, that one of the few Republicans with balls, as it were, is a woman. I salute Liz and hope to see and hear more from her in the future (as I have said before, America is rotting in many respects and I would like to see that process reversed somehow, but I would hardly follow a parody of a man to accomplish that goal, so cannot support the last Trump).

And speaking of the herd aimlessly moving this way and that, a few weeks ago the major media all tried to shift from coverage of the courageous fight of the Ukrainians against the murderous, cowardly hordes of Putin, to the plight of the Afghan people under the Taliban. There was little mention of the twenty years, 3000 or so American lives, and trillions of American dollars spent giving those people every chance to achieve a democratic republic and dispose of the Taliban. Apparently we are somehow supposed to believe it is our fault. I suppose the the pc/multicultural/antiwhite/antiwestern axis is nervous that too much attention is devoted to the Christian, white, courageous Ukranians. In any case, that dog didn't hunt, as LBJ might have said, i.e., the media dropped most of that new focus as the public found it unappealing and unbelievable.

New product idea I am considering: Energy saver thermostats for home HVAC, simply adds a degree to any setting for cooling and subtracts a degree from any setting for heat. 

Well-done, Speaker Pelosi (another case of 'nads). You thoroughly annoyed the Chinese Totalitarian Club, but at least made it more likely that we would fight the Chinese if they attacked Taiwan. Why should we care? Aside from our word as a nation, you must judge from the lessons of history that totalist regimes never stop with their immediate neighbors. To that extent, better to help others fight than find ourselves isolated facing a fight for survival alone. Of course, with our population increasingly stupid and obese to a degree that our Armed Forces our having difficulty finding qualified recruits, any fight we get into is going to be a tough one now. We still have some technological superiority, but that won't last forever (American science is slipping, not surprisingly when concepts of excellence have been discarded in favor of diversity for its own sake and education has been made accessible to all and therefore of little worth). 

Good night, and good luck.

No comments: