Friday, November 30, 2018

I Need Some Time to Review

There has been media attention for a report from the federal government about the possible effects of global warming. Some of the attention has focused on the president's dismissal of the report. He claimed very high intelligence, but said he isn't a believer. In the context, he seemed to be saying that he doesn't believe in much of anything, because there is something bad about belief, or some other nonsensical idea. He certainly didn't address the data or the conclusions of the report. I have my doubts about his intelligence. I don't think he is capable of reading the report and understanding it well enough to judge the conclusions, nor do I think he has any interest in making the attempt. He spouts whatever occurs to him, and gets angry if anyone asks questions, even if the questions are just due to confusion because what he says makes no sense to anyone but him.

But I did read an analysis of the report that seemed a little more substantive, and I feel like I should follow it up. The author said that the report predicted a decrease of 10% in economic growth, which other sources have also reported, but he said that was against a baseline of 300% growth by 2100, meaning our economy would drop to 290% total growth over that period with global warming factored in, if, and only if, the warming was near the maximum level in the predicted range. He goes on to say that most of the loss of growth is predicted to be due to an increase in deaths from extreme heat. He then claims that the level of warming necessary to drive this is much more than is actually predicted by the current models, and that the deaths would be easily avoided anyway.

Then he compares the effects of proposed plans to mitigate global warming, starting with a modest level, which would cost some $10 trillion, but is predicted to pay for itself, plus a little bonus, over the same period of time, and then looking at plans to hold global warming to 1.5 degrees C, as per the Paris Accords, which he says would cost some $140 trillion, and would have less than a tenth of that in benefits.

Most of the articles I've seen do little to put the predictions of the report in any context, so I can't say whether this author is giving an accurate portrayal of the report, or the projected costs, which it didn't seem all came from the report. There is a lot of uncertainty in something as complex as extrapolating global warming for 80 years, and including economic effects. It is difficult just to make accurate predictions of economic growth for 80 years, without factoring in something like a fundamental change in climate. But that is what is called for in this report, and in the process of making decisions about policy regarding global warming. I think we may be past the point of choosing modest measures, like pollution control, that would have clear benefits on health, and by extension, economic activity. I think we need to try to predict sea level rise, and how it will affect island nations and coastal areas. And what about northward migration of animals and plants? And changes in growing seasons, and milder winters, with increased insect survival, and all kinds of other factors that go into any kind of predictions of economic effects over decades.

So, I will try to track down some other information about the report. I may even have to read it, in the end. I could ask the algorithms to do this analysis for me, but they still can't figure out what I might and absolutely won't be shopping for in the next year, so I have no confidence in them.

Sunday, November 25, 2018

Endurance of Good Ideas

I have just finished reading, or perhaps rereading, Darwin's, "On the Origin of Species". I thought I had read it before, but there are things I think I should have remembered that I noticed this time. There was a lot that Darwin didn't know.

I have about decided, though, that the next time someone starts a debate with me over the Theory of Evolution, I will insist that she read Darwin first. Most of the people who have made effort to convince me of the errors of my ways in "believing in evolution" would have a very hard time reading Darwin. His breadth of knowledge in the natural sciences was impressive, and his writing is rather dense and has big words in it. He writes like I think, sometimes.

Darwin was writing a hundred and fifty years ago. The edition I read was the last one, I think, and included chapters responding to criticisms of his earlier editions. A hundred and fifty years ago, there was debate over the age of the earth. Some would say there still is, but that is not true among geologists and other scientists, at least in the main stream of thought. But in Darwin's day, the majority even among naturalists hadn't thought much about the age of the earth, and kind of just accepted the biblical history uncritically. Darwin knew that his idea of natural selection required much more than six thousand years of biology to account for modern species. Some geologists and astronomers were coming out with theories about the age of the earth based on thermodynamics that extended that age into the millions of years, which helped Darwin, but he realized it was still probably not enough. Lord Kelvin (who hadn't been given the title yet) calculated a rate of cooling for the earth from a mass of liquid lava to the condition in the 1800's, and estimated the earth to be 20 million years old.

Darwin speculated, correctly, that the earth was older than that, because he saw evolution from natural selection as being very slow. If the rate of variation and selection could be estimated, he said, then time could be estimated even in the fossil record based on the changes seen in species in different layers of rock.

The big change in estimates of the age of the earth came from the discovery of radiation, and the contribution radioactive decay made to the energy balance and cooling of the earth. The current estimate is 4.54 billion years. And the fossil record has filled in a lot since Darwin's day, confirming his predictions.

I was a little surprised, though thinking back I should not have been, to see that in 1859, the distribution of electrons, which explains the organization of the periodic table of the elements, hadn't been discovered. That came from the Rutherford nuclear model of the atom, and the Bohr model of electron configuration in the first decades of the 20th century.

It is rather remarkable that Darwin got so much right, based on his limited data set. Okay, he had studied thousands of species, at least to some extent, and was familiar with detailed descriptions, including careful measurements of various parts of plants and animals. He knew about the debates over classification of species and varieties, and the patterns of distribution of a lot of species. He found the necessary patterns to support his ideas, with help from other naturalists, including Alfred Wallace, who actually wrote a paper about natural selection before Darwin had published his own. But he had no idea about the source of variation, at least at the cellular and molecular levels. That took another hundred years. It pretty well nailed down Darwin's theory, though. Now we can trace genealogies of species in changes in DNA. And we have. That makes it hard to argue honestly against common descent, but there are those who still claim Darwin was wrong. Usually, they ignore what Darwin really said, or ignore the obvious consequences of acknowledging the observable facts.

So, having read Darwin, I am convinced that I have been right all along to believe in the Theory of Evolution by Natural Selection.

I do not know how quickly modern changes, both by natural and technological mechanisms, will occur, or what directions they will take. I have still not seen much cause to be afraid of the Rise of the Algorithms. They still send me ads for colleges and programs, as if I were shopping for programs to pursue another degree. I am content with my formal education. I will continue to read and think and learn, but I don't know about taking classes. After all, I teach college. I also get a lot of ads for medications I don't need. If I develop a disease that can be treated by these medications, I may start thinking differently, but for now, the algorithms seem completely off in their analysis of what might catch my attention.

If I tell them there is no history of diabetes mellitus in my family, and I have two doctoral degrees, will they change the ads I see? If I say I like reading science fiction and fantasy, will I get ads for those? Hah. The algorithms are pathetic weaklings, and I expect they will still be when I stop caring.

Saturday, November 10, 2018

A Look at the Data in the Election

I was wrong about vote totals in Ohio. Republicans did get more votes in the races for congressional representatives. I think the final count was about 51.5% for Republicans and maybe 46.5% for Democrats, with 2% for other parties. All the statewide offices except U. S. Senate were won by Republicans. Sherrod Brown won re-election to the Senate, but did not seem to pull many votes into other races for Democrats.

One thing I noticed in reading through the vote counts. Closer races had more voters. Many of the congressional districts in Ohio are quite skewed to one party, four for Democrats and twelve for Republicans. The less skewed districts had more voters show up. Since the four districts set up for Democrats to win are very skewed in favor of Democrats, those races had the fewest voters participating. I have to wonder if that is bad for democracy.

I'm not the only won who wonders about that. Enough people have expressed concern that there was a referendum on the ballot in the primary elections in May, and a strong majority voted to change the process for setting up the districts, requiring a bipartisan committee to make the new districts. If votes from the minority party are not gained for the plan, a new plan must be created in a few years, rather than waiting for the next census. I don't know if that plan will result in more competitive districts, but it should be better than the current plan, which allows the majority party in the state legislature to set all districts, which is what created this system in which an election with a difference of a few percent always results in the same distribution among elected representatives.

My district, for example, had a shift in voting 13% in favor of Democrats in this election, and the Republican still won re-election easily. All but two districts in Ohio shifted in favor of Democrats. The two that shifted toward Republicans were both majority-Democrat districts, which were still won by Democrats. All the majority-Republican districts shifted toward Democrats, ranging from 3% to 32%, with most in the teens. It is hard to say what that will mean in coming elections. A shift of 10% toward Democrats in the 2020 election compared to 2016 would mean more Democrats in the state legislature and a different result for president. It might mean a state house majority, but not a state senate majority for Democrats. Sadly, I don't think it would make a difference in the makeup of the Congressional delegation. A shift of another 10%, meaning a total of 20% from 2016 results, would mean a big change, with eight republicans losing, but it would take almost that to change the outcome of more than two races.

Of course, elections in presidential years always have greater voter participation than off-year elections. In some states, there was a large increase in voter participation for this midterm, but not so much in Ohio. There was an increase, but the voter turnout was still below 50%, and well below the 70% in presidential elections. This year looks about 1% below the turnout in 2010.

The results seem to be within the margin of error of the polls, but instead of Democrats winning statewide elections by a percent or two, Republicans won by a percent or two. Okay, maybe 4%, but a shift of 2% from Republican to Democrat would have turned the races. I based my prediction on the polls and the predictions that the Democratic candidate for governor would win by a percent or two. I figured that would pull along some votes in the congressional races. That didn't happen.

There may be a small, but important systematic error in polling in favor of Democrats here. I have wondered about the accuracy of polls with the rise of cell phones and so forth. It is hard to know whether the samples used are representative. I keep hoping the polling professionals know what they are doing. The general outcome of the election nationwide seems to be within the margin of error of the polls and predictions, so that is good, but the variances still bug me. I really wanted Richard Cordray to win.

Sunday, November 4, 2018

A Paucity of Polls

In an earlier post, I made a prediction about the midterm election in Ohio. Today, I looked for recent polling data, to see if there is any relevant information that might lead me to update my prediction. There isn't much.

There are some polls about the Senate race and the governor's race in Ohio. In both, there is a sense that the Democratic candidate has a small lead in polls, but around the size of the margin of error. I could not find any data on most of the Congressional races, though there are a couple of polls since the special election in August about the 12th district. That seems like a close race, based on the polls, but FiveThirtyEight still rates it as lean Republican, based on historical data for the district.

I also did not find any data on expected voter turnout or early voting trends in Ohio. Maybe I'm the only one interested. After all, the statewide races in Ohio have drawn some national interest, as Ohio is still considered a bellwether state in national elections, and having a Democratic governor in the next presidential election gives the Democratic candidate a small advantage over having a Republican governor. But, with the possible exception of the 12th district this time, the results of the congressional elections in Ohio are hardly in doubt. It would take a substantial blue wave to overcome the gerrymandering here. The current national generic congressional polls show an advantage for Democrats of 7%, which is not enough to overcome the built-in advantage of Republicans in any of their 12 majority districts.

That is that state of democracy in Ohio. A 7% shift over baseline may help elect a Democrat for governor, and an incumbent Democratic senator may easily win re-election, but Congress won't be affected. Democrats would need another 3 or 4 percent shift to win more of those races, except in the one "open" seat. My question was whether Ohio was in line with the rest of the country in the generic congressional ballot polls in the size of the shift. I could not find any data to tell me.

So I will stick with my original prediction. Democrats will win some of the statewide races, maybe a lot of them, and will get more votes than Republicans for congressional seats in total, but will still only win four (or maybe five) seats out of sixteen.

What happens to Ohio and the country after that, I feel unable to predict. We have a president who appears to be using the military for political theater, and who has told some astounding recent lies about tax cuts and the elimination of birthright citizenship in what seems to be a desperate attempt to stave off substantial losses for Republicans in Congress, and some of his followers are believing him. He has also said there would be violence in Democrats won. And now, nationally, it looks likely Democrats will win enough seats in the House to take control. I cannot predict the response of this childish president or the worst of his followers, but I worry about it. How will he deal with the label of loser?

I did find some data on early voting in Ohio, which is not so remarkable as in many other states. There is a list of states with more early voting to date than was seen in the last midterm election, and some, including Texas, which have had more early votes than the final totals in 2014. Not Ohio. As of 30 October, Ohio has had maybe a hundred thousand fewer early votes than in 2014, which I think puts it ahead of the same time before the election last time, but only by a little. There may be some increase in early votes and total voter turnout compared to recent midterm elections, but there does not appear to be the same excitement in Ohio as in many states.

Maybe that is because the senate race is not very close and will not mean a change. The governor's race doesn't seem to be generating a lot of excitement, either. I don't know why. I think Richard Cordray is a very capable man, and will make a very good governor. I do not have the same thoughts about Mike DeWine, who holds rather unappetizing views. As a senator, he supported a Flag Protection Amendment to the Constitution, which is essentially a welfare bill for trial lawyers, and would end up wasting a lot of federal court time over arguments of how red red is, and exactly what number and width of strips of cloth must be protected. Yes, there is such a thing as a stupid question. We are all better off for Senator DeWine's failure to get that bill passed. His views on other issues are equally ill-conceived, in my experience. I think he should retire from public life, and let more thoughtful people take over.

But the voters of Ohio don't seem to care very much, compared to the counterparts in other states. We'll see what the final tally is. There has not been an update on the early vote numbers since last Tuesday, so there may be a substantial increase in early voting by the time all the votes are in. After all, absentee ballots can be counted if post-marked on the day of the election, so they could be arriving in numbers for a few more days. But the number requested is only a modest increase from 2014.

What would it take to get Ohio voters excited? I'm not sure I want to know. I have seen some of what Ohio voters choose, and sometimes, I just shake my head in wonder. What I'd really like to know is how to excite those who would vote like me without disturbing the others, leaving them to sit quietly at home while my co-believers made our statements and tried to make the world better. I am not a political scientist nor an activist, however, so I just think about how nice it could be, and wait for the actual election results.

Thursday, November 1, 2018

Poor Predictions

I haven't had much cause to think about the rising power of algorithms recently. Maybe I've settled into a comfortable pattern, and just missed the actual control those sneaky electronic devils have over my life, but I really haven't seen much.

My wife and I listened to, "Ready Player One", read by Wil Wheaton. I enjoyed it. Sure, it's a post-apocalyptic tale of young people with no future spending their lives playing video games, battling a huge and well-financed corporation over the future of an online virtual reality system, so no one ever addresses any of the real problems facing the human race in their time, instead opting to spend all their time and energy on entertainment, consuming resources and doing nothing of any value, but it was  fun. The poor, suffering teens overcame all the odds, and solved the puzzles that were beyond the abilities of the world's finest minds and supercomputers because it was about '80's culture, which not logic can interpret, and to which reason should never be applied. But some of the artificial intelligence was impressive, and the hero took a lot of chances and found clever ways to do what he needed to do to win.

But it didn't make me fear the algorithms. It was set around 2045, and the world was dying because of global warming and an end to the supplies of fossil fuels. But we have better solar technology than they did already. If the multi-billion dollar corporation had spent its money on solar power instead of oology, it could have saved half the planet instead of just losing to a high school dropout.

Will we follow the path that leads to dystopia? Our current president seems to want to. He is trying to increase extraction of fossil fuels, auctioning off leases of public lands for fracking, and eliminate restrictions on emissions, so we can add more and more carbon to the atmosphere. Short term economic gains at the expense of a polluted and warming world. I think even our current artificial intelligence knows better, to say nothing of actual science.

Of course, this president has also said that if his party loses control of Congress, the stock market will crash, and the dystopian future we all fear will come next week.

I really hope the electorate decides that's a theory worth testing.