Groundhog's Day and the Meaning of Life

Yesterday was Groundhog's Day, the holiday where everyone waits with baited breath for a rodent to decide if it saw its shadow, and therefore, whether there will be a long winter, or an early spring. 

I don't actually know what happened this year.

But to celebrate the auspicious occasion, we went to my in law's house, and we watched the movie "Groundhog Day" with Bill Murray. 

I had forgotten just how great (and even Buddhist) that movie is. 

Bill Murray plays Phil Connors, who is quite the self centered jerk, is a reporter sent to cover the groundhog’s day celebration (a holiday that he thinks is ridiculously silly and pointless), but for some unexplained reason, he begins to relive that frustrating day over and over again, forever. 

What made the movie especially interesting to me this time through is that Phil seemed to move through many of the stages of an existential crisis. The movie is primarily a comedy, but I think it has a lot to say about finding meaning and purpose in life. Bear with me here.

When Phil first realizes that his actions have no consequences, he immediately experiments with Hedonistic gluttony, then moves through the blind pursuit of sexual pleasure, to greed as he learns to steal with impunity (because he can try over and over again until it works, with no consequences for his failures). 

But rather predictably, this doesn't make him happy and Phil becomes depressed. What do you do with an infinity of life? Shallow pleasures are simply not fulfilling when you live the same day over and over again. This is the existential horror expressed by Hindu’s and Buddhists at the thought of Samsara and eternal reincarnation. What good is eternal life anyway? One option of course, is suicide. In this context, why not simply end the cycle of rebirth? But it doesn’t work. To his horror, he wakes up in the same bed, with the same horrible morning alarm music over and over again. No matter what he does. He’s trapped in Samsara, the eternal cycle of rebirth forever. 

But the next stage of his existential crisis involved his falling in love. That could have provided a new sense of meaning for Phil, except he can’t win over the woman he loves in a single day. She remembers his past behavior towards her, and he has to start over with any progress the next day. Worse, he comes to realize that in his selfishness, he is simply not worthy of her. 

And this is when things begin to change… and he begins a process of self improvement. He studies French Poetry. He initially despised it, and only learned about it to seduce his love interest, but eventually he finds that he likes it. He takes piano lessons. He learns to ice sculpt. And then, finally, he begins to live for others. This is the key. 

Knowing what will happen each day, he starts to intervene in ways that make the lives of others better. Instead of mocking the celebration of groundhog’s day, he finds a way to make it inspiring for the other people who enjoy it. He changes a flat tire, he catches a kid who falls from a tree. He tries to save the life of an older homeless man. He saves another man from choking. Then, while dancing with his love interest, he says that whatever comes next, he is “happy now”. He has found a way to live for the moment, and find joy IN that moment, even if he can never escape the eternal cycle, he’s happy. 

And this is why he was wrong that having to relive a day means that there are no consequences for his actions. Harmful actions are wrong precisely because they cause harm to others in the moment, regardless of what comes after. Similarly, good actions are good because they cause joy for ourselves and others in that moment. It's true that the future matters as well, and that actions that cause joy now but suffering later can be wrong because of the future effects. But this does not mean that the lack of a future effect robs the present joy or suffering of all meaning. 

Compassion, love, and living in the moment have removed his suffering, and made a bad man into a good one. 

And now, finally, he is worthy of the woman he loves, and when he falls asleep in her arms, he wakes up the next day, and he is finally free of Samsara. It’s February 3rd, and groundhog’s day is over. But then, like the Bodhisattva, he says that he wants to live in the town he tried so hard to escape from. He loves the people there now. All of them. 

And now his life has meaning. 

I think this is a rather profound illustration of my own beliefs, and the way that I find meaning and purpose in life, without any need for belief in a god, or heaven, etc. Life is beautiful, because we can be happy, now, in the midst of Samsara. And we can choose to return to Samsara, rather than just seeking an escape from it, because we can find meaning in living a life for others, helping them to escape their own suffering, and finding joy in each moment. I think that compassion, love, relationships, are what makes my life worth living.

I think that Groundhog's day has joined with several other key movies/stories that I believe function as "modern myths", in that they explore or illustrate concepts around the meaning of life in a profound and moving way.  

Spurious Claims of Voter Fraud

Because of the absolute blizzard of misinformation and flat out lies being currently spread, it is essential to share the facts with regard to Trump's dangerous, authoritarian, and undemocratic claims of voter fraud. I have done my best to provide my followers with accurate, factual, and up to date information, based on data and evidence during this entire process. It is disappointing that I even have to address this issue, but I apparently do. 

I grew up hearing spurious claims about "dead people voting" etc. I heard these claims my entire life, and I often believed them, without actually looking into the evidence for them. We need to understand what constitutes good evidence, and the ways in which claims like these should be tested. The fact that things like this are said on talk radio, is not proof that they are accurate and reliable. Since then, I have actually taken an interest in voting, polling, statistics, and the data science behind it. In all that work, I have found that the evidence for these claims, beyond simple claims and hearsay, is practically non-existent. It always was. 

After intense claims of voter fraud in 2016, Trump initiated an investigation and task force to look into all those claims, the so called "Voter Integrity Commission". Pence was in charge of this commission/task force. It concluded with no concrete evidence having been found or presented to the American public. https://apnews.com/article/f5f6a73b2af546ee97816bb35e82c18d The burden of proof continues to be on those claiming fraud, and our courts of law are the places where such evidence should be weighed and evaluated. 

As Lyle Schofield, a dedicated (and frankly extreme) Republican and Trump supporter wrote: "if [Trump] understood world history and geopolitical reality he is playing right into Russia, China, and Iran's hands by creating distrust in our form of government and the outcome of the election." https://www.facebook.com/lyle.schofield.7/posts/3586941988015270 Even he sees it. 

Romney recently responded to Trump's claims by pointing out that Trump "has a relatively relaxed relationship with the truth." https://twitter.com/CNNPolitics/status/1325448775595339779?s=04 

The American people are well aware of the fact that Trump often says whatever he wants, and thinks that if he says it often enough, people will think it's true. Only 39% of those who voted said that they thought that Trump was "honest and trustworthy". https://www.npr.org/2020/11/03/929478378/understanding-the-2020-electorate-ap-votecast-survey

Republicans USED to care about these geopolitical issues. The collateral damage to our democratic institutions and international standing may be irreparable.  

For this reason, there should be a high bar of evidence for such claims... A bar that Trump and his team are clearly not crossing.

The Court Battles, Evidence of Fraud: 

Litigation regarding which votes should be counted, and how they should be counted, is common in elections like this. That part at least is somewhat normal, even if the rhetoric around it this year is not. This litigation leaves a paper trail of actual, reliable, confirmed evidence. 

The best way to assess the strength of the claims of fraud is to simply look at the results and arguments on his ongoing court battles. Unlike at a press conference, in court you have to have (and present) factual evidence. Press conferences, social media, network news anchors, are all irrelevant. Simply read the court filings, and follow the rulings of the various judges. Remember, the courts are largely packed with Republican judges, they are NOT somehow biased against Trump. Even if you believe that the media is biased against Trump, the courts are not. 

He has every opportunity to get a fair hearing there. And yet, none of these cases are actually going well for Trump.

Even more telling, most of the cases don't even allege what they are falsely claiming in their press conference, because in a court of law, you can't simply make things up like you can in a press conference... you need actual evidence, and they apparently have none. If they did, they would be presenting it to the courts. And they are not. 

Here is a useful summary of the current state of the various court cases, both those that are finished, and those still in play: https://www.washingtonpost.com/politics/trump-election-irregularities-claims/2020/11/08/8f704e6c-2141-11eb-ba21-f2f001f0554b_story.html 

As Chris Christie (Republican) has repeatedly said, Trump should now provide concrete and court worthy EVIDENCE for voter fraud, or move on. "That's why, to me, I think it was so important early on to say to the president: If your basis for not conceding is that there was voter fraud, then show us. Show us, because if you can't show us we can't do this. We can't back you blindly without evidence." https://nj1015.com/christie-on-trump-provide-proof-of-election-fraud-or-just-move-on/

Republicans like Christie understand the incredible damage that will come to our democracy if Trump eventually concedes that the election was "stolen" rather than that he simply lost. 

If there is actual, concrete, and actionable evidence of fraud, it should be heard by the courts, and the vote counts corrected appropriately. But so far, none has been found. If any is found, I will be the first to call for the correction of the vote counts. 

The Red/Blue Mirage, Evidence for Fraud? 

Another very VERY essential point is that the most intuitive argument for fraud being put forward is that Trump was INCREDIBLY far ahead, and then fell behind as the slower mail in votes were counted. 

Here I can say something relatively concrete, because we are back in my area of expertise (polling, uncertainty quantification, and data analysis). 

Before the election began, I pointed out that this year we should expect wild swings depending on when each state counted the different TYPES of votes. I reminded people of this on November 3rt, ad 8:33pm Mountain:

"Reminder, we expect a late red shift in Arizona, but a blue shift in Wisconsin, Michigan, and Pennsylvania... The difference has to do with whether mail-in votes are counted and reported first... or after..." https://www.facebook.com/jlcarroll/posts/10160290376544018: 

We knew this going in. For example, Ohio looked VERY good for Biden, but then swung hard towards Trump. Why? Because Ohio counted the mail in ballots first BEFORE election night even began.

Pennsylvania wanted to do the same, but they were blocked from doing this, and could not begin processing mail in ballots until election night. The result was that in person votes were reported FIRST, while mail in votes were counted second. 

Here's what the AP Votecast "alternative" to exit polls shows about mail in voting:

​

https://www.npr.org/2020/11/03/929478378/understanding-the-2020-electorate-ap-votecast-survey 

The reliability of exit polling this year, given the errors in the pre-election polls, and (more significantly for exit polls) the amount of mail in voting this year, is another topic of discussion that I hope to write about soon. 

But for now, it's enough to say that our best guess is that Biden won mail in voting by 36 points, more in other regions. This is not surprising given that one candidate discouraged his followers from voting by mail, and given that the COVID-19 outbreak has been politicized such that one party is more concerned about it than the other. 

I KNEW this late shift in counts was a likely thing, and I did my best to warn everyone before the election began. And yet, even I was confused about it twice during election night, once when I started to think that Biden had a shot at Ohio, and again when I started to think that Trump was more likely than Biden to win Pennsylvania. In both cases, I KNEW the late shift red in Ohio, and blue in Pennsylvania was coming, but I questioned whether it would be enough.  During the course of the election, I NEVER thought that Trump was the inevitable winner, but I did think he was up.

In contrast, fiveThirtyEight's conditional probability model (taking the called states into account) kept telling me that Biden was a large favorite to win Michigan and Wisconsin, and that Pennsylvania would be close. I should have trusted it more. It's also true that there were election experts who were looking at which types of votes were still out in Pennsylvania, and where those votes were from, who were telling me at the time that Biden was actually up in Pennsylvania the entire time, even when I was saying that Trump was likely a small favorite. This was simply a matter of lacking information on my part. In retrospect, I can look at the raw data, and see that a Biden victory in Pennsylvania was almost assured from very early on on election night. Trump was NEVER actually ahead there, given which votes were in, where they came from, and what kind they were.

The point of all of this is to say that there is no evidence of voter fraud in a late blue, or red shift in these states. It was expected, we understand the mechanism, and it is not at all surprising, or nefarious.

The math simply doesn't support that claim, so if there is actual evidence of fraud, it will need to come from another source... But as I said above, so far, the court battles haven't provided it. 

COVID-19 Daily Updates Moved:

New Blog for COVID-19 Daily Updates:

I have been posting my daily COVID-19 updates here, on my personal blog. But it now makes sense for this to have its own space. So I will be posting all future daily updates there. If you have been following me here for those updates, I recommend subscribing over there: 

https://covid-19watch.blogspot.com/

COVID-19 Daily Update for Yesterday, Sunday, 8/30/2020

World Wide Trends, How Many People Have Been or Currently Are Infected? 

There have now been 25,417,886 confirmed cases of COVID-19 world wide.  That's 0.33% of the world population. According to WorldOMeter, 6,840,933 of those cases are "active". However, as we discussed in the last update, WorldOMeter's recovery estimate is too low. I estimate that there are currently 4,395,451 confirmed cases that are still "active'. 

However, how many people have actually been infected? Many people who are infected are never tested, and never become a "confirmed" case. 

One way to estimate infections that I have used before, is to use the number of deaths, and the range of possible infection fatality rates to get a range of possible true infections (and assuming an average 20 day lag between infection and death).

This estimate will only be as good as the approximations to the Infection Fatality Rate (IFR), and the reliability of the death data used. I am currently using the range of IFR estimates given by the CDC. 

If we apply this to the World-Wide data, these are the results: 

We can see that the number of true infections inferred in this way is significantly larger than the number of confirmed infections. But it's also worth pointing out that this is possibly an under estimate, because it is highly likely that the true death tole world wide is under-reported, especially in less developed regions. 

We can then apply the same formula we used last week for estimating recoveries to turn the cumulative number infected into an estimate of the number of active cases, those currently infected: 

By this estimate somewhere between 0.15% and 0.3% of the world's population are currently infected with SARS-CoV-2, the virus that causes COVID-19. 

Both of these estimates seem unreasonably low. Certainly it is significantly lower than we see in Europe or the US (although the US should hardly be held up as a model of success). Certainly a large portion of the world's population live in China, and they have done remarkably well at limiting the virus' spread. 

Nevertheless, it seems likely to me that the true death tole from COVID-19 is significantly under-reported world wide. 

Deaths By Day, Estimating the Reporting Lag from Florida:

It can be very helpful when locations report either cases by date of symptom onset, or deaths by date of death, rather than by the date the case or death was reported/recorded. This can give us a better idea of the true shape of the curve... but it also has a lag to it, that always makes it look like cases/deaths are going down while we wait for these cases/deaths to be reported. 

When modeling this lag, it's important to have good data on what the lag has been historically. For Florida, there is an excellent archive of historical information showing how the data has come in over time, available here: 

https://github.com/mbevand/florida-covid19-deaths-by-day 

This is what that delay looks like for Florida: 

The proposed fit is 1-e^(-0.1513x).

COVID-19 Daily Update for Yesterday, Friday, 8/28/2020

World Wide Trends, Updated Methodology:

I have been using WorldOMeter to get world wide recovery data. They had the best estimates, because many places simply don't report recoveries, and in that case, WorldOMeter seemed to be estimating them. 

Sometimes. 

But after staring at this figure for weeks now, I'm coming to the conclusion that they are missing a lot of recoveries. 

It's been over a month since peak cases back in mid July, but the recoveries have not yet caught up. But they should have by now. And if this was wrong, it would be impacting everything I did with global trends... the % growth rate, doubling times, etc.

So just as a sanity check, I implemented a common "recovery estimation" algorithm (used for a while by Texas). That is, I took the number of cases-deaths, and I assumed that 80% of those will have a mild case, and will recover in 14 days, while 20% of those will have a more severe case and recover in 32 days. 

(Note, I am not looking for "long haulers" who seem to have a lingering immune over-reaction that persists and causes sometimes debilitating problems long aver the virus itself has left their system. Rather, I am looking for the number of people who have cleared the virus and are therefore no longer contagious, regardless of long term impacts.) 

As you can see, this approach paints a far more optimistic picture of what is happening. I also think that it is more likely to be correct. 

If we use these estimates of recoveries to estimate the daily change in active cases, we get the following: 

Again we see that my estimates look very different from those of WorldOMeter, and, they paint a much more optimistic picture of what is happening. Right now, my estimate shows that the number of active cases is declining (the change is below 0) while WoM's estimate for recoveries has the number of active infections still growing (above 0).

If we use these new recovery estimates to calculate the % growth (in active cases) this is the result: 

And if we zoom in on the y axis we see that the trend has been a negative % growth for just over a week: 

I highly suspect that right now more people are recovering each day than are being newly infected each day. Which we would not suspect if we just naively used WoM's recovery data.

Some good news. 

CODID-19 Daily Update for Yesterday, Thursday, 8/27/2020

World Wide Trends in Brief:

The world wide daily change in new cases now appears to be roughly flat in both the 7 and the 14 day averages: 

The reason for this can be seen in the daily cases and recoveries chart below: 

However, since the daily cases are also in VERY gradual decline, eventually the daily recoveries should catch up, and the daily change in active cases should go negative. Unless there is a new increase in the number of daily cases.

Second Wave Watch:

For the last month we have been watching three countries that initially did very well in responding to the virus, namely South Korea, New Zealand, and China. They each used a different approach, but all three initially succeeded in handling the pandemic in a way that most of the rest of the world has failed to do. New Zealand even completely eliminated all local transmission for over 100 days, and completely reopened. 

However, near the end of July/start of August, all three have seen a second resurgence of cases. It will be important to watch and see if they are capable of controlling the virus again, if so, it will demonstrate that their initial success was not a 'fluke', and that it was something the rest of us COULD have done, had we chosen to. 

Cases in South Korea are still rising. Their contact tracing system has yet to catch up with the viral spread. Nevertheless, most countries would be ecstatic if they could keep cases below 400/day. Whether this remains a "success" will depend on whether cases continue to climb. 

Cases in New Zealand appear to be declining. Their response to this failure at their boarder which allowed the virus back in has been phenomenal so far. If this continues, it seems likely that New Zealand will soon be virus free again, and can again reopen. 

When they do, they will show the world that the best response both economically and from the perspective of public health is to "go early, and go hard."

China is also seeing a steep decline in the number of daily cases, which is now well below 100/day. 

While their authoritarian tactics are neither desirable nor something we would want to emulate, they are a second case study demonstrating that local lockdowns coupled with contact tracing can successfully eliminate viral spread, and keep the rest of the country open, and thus minimizing economic impact. 

And New Zealand demonstrates that this can be done without the authoritarian overtones. 

The US's halfhearted response, with a "lockdown" that never actually locked things down, is the absolute worst of both worlds, with extreme economic impact, but without any meaningful control of the virus. 

COVID-19 Daily Update for Yesterday, 8/26/2020

Automation Update, Available Figures:

I am now automatically generating several figures each day and pushing them to an online "git" repository. I have had two volunteers offer to work on creating a web page that will provide a more useful interface to the figures/data. But for now, you can browse the figures with this link. 

There are several things available so far. 

First, I am running the rt.live algorithm on the US as a whole (which they don't do). There are some good reasons they don't do that (the US is a diverse place, and some areas are doing better than others). But for those interested in what their algorithm has to say about the US as a whole, I now provide that visual updated every day.

As I have previously discussed, I disagree with the conclusions of rt.live's algorithm. I believe they are FAR too aggressive in adjusting for testing rates, which causes them to assume that the first peak(in April) was much higher than the second (in July). I believe they were of roughly the same size. So I hope to eventually provide a modified version of this analysis that more accurately adjusts for testing rates. 

We are providing tests and cases for each state. As an example, here's New Mexico: 

For these figures, the scale is set such that the tests are at a scale 10 times that of the cases, therefore, if the cases are above the tests in the figure, then the percent of tests that are positive is over the danger zone of 10%. The goal is to keep the cases (orange line) well below the tests (blue line).

 

And finally, we are providing the % of tests that are positive information directly in its own figure. As an example, here is New Mexico again: 

Again, the goal is to keep the % positive rate below the danger zone of 10%. 

Because I am now producing these figures for each state every day, you can check your own state, without waiting for me to do one of my "daily updates" on the state you care about. 

This is all a work in progress, so hopefully the user interface for viewing all this data will improve substantially in the future.