# Predictions and probabilities from Professor Ceiling Cat

April 21, 2020 • 8:30 am

Here are several items for which I’m willing to assign Bayesian probabilities, updated in light of the most recent information. Please join me and either revise these to your own specifications, or make additional predictions. These are not necessarily what I wish, but simply what I think will happen. And of course I’m just guessing here, rounding to the nearest 5%.

Probability that Kim Jong-un is in serious medical trouble:  65% (probability of death in next several weeks: 50%)

Probability that schools and universities in the Midwest and Northeast U.S. will reopen in the fall:  45% (we will have no vaccine and schools mean crowding)

Probability that schools and universities in the Deep South will reopen in the fall: 75% for schools, 50% for universities

Probability that we’ll have even a partial Major League Baseball Season: 40%

Probability that Trump will be reelected in November: 35% (as you know, I’ve bet \$50 that he’ll lose)

Probability that students will learn even half as much under remote instruction as they would with in-person learning:  5%

Probability that I will see an ad on television today telling us that “we’re all in this together” (as if that helps): 100%

Probability that an average American can correctly draw a “regular” (no intervention) versus a “flattened” pandemic curve, with the axes properly labeled: 20%

I also predict that Honey will hatch eight ducklings.  That’s the size of her brood two years ago (two had been lost by the time of this photo, but all of these survived and fledged):

## 73 thoughts on “Predictions and probabilities from Professor Ceiling Cat”

1. ThyroidPlanet says:

Interesting…

2. GBJames says:

I can’t disagree with any of those probabilities other than the one about “We’re all in this together” on TV at 100%. In my case the probability of watching commercials on the TV machine is probably only about 10%, so the chance of me seeing such a commercial is quite low.

1. infiniteimprobabilit says:

Oh, well said. That goes for me too.

(The probability of me ever watching TV commercials, that is. I’m in NZ so I wouldn’t see those specific commercials anyway, but we have our equivalents of course.

In fact I have a good DVD collection I’m working my way through again so the TV rarely gets switched to broadcast channels).

cr

3. enl says:

Can’t argue with any of these.

4. Charles A Sawicki says:

“Probability that Kim Jong-un is in serious medical trouble: 65% (probability of death in next several weeks: 50%)”
I don’t see the support for this inference of his short life. I bet that he will be alive in October.

“Probability that Trump will be reelected in November: 35% (as you know, I’ve bet \$50 that he’ll lose)”
I hope you are right about this one, but a pandemic makes possible so many ways to reduce voter turnout in those crowed cites full of Democratic voters. The less populated areas where it is “safe” to vote are mostly strongly Republican.

“Probability that students will learn even half as much under remote instruction as they would with in-person learning: 5%”
I agree on average, but smart motivated students may actually learn more. Much of school is boringly repetitive.

Otherwise I agree with your predictions.

5. Unless Trump succumbs to corona itself, he will be re-elected (my guess >85%).

5% is right for online teaching. My kids are learning far less but they and other kids have not turned their brains off. We are still watching science, getting exercise, and talking parodies, creating art.

1. paablo says:

Agree about Trump. We’ve already had several chances to refute Trumpism. Where is the evidence that suggests people will actually turn out to vote?

1. Saul Sorrell-Till says:

What do you mean by “refute Trumpism”? It seems to refute itself.

2. Ken Kukec says:

What’s the basis for your Bayesian priors of >85% on a Trump victory, Kevin?

I take it you’re aware of the current polling data and of Trump’s miserable approval ratings over the course of his entire presidency?

Over 85% seems awfully high for a candidate who has essentially no chance at all of winning over 50% of the popular vote (and who’s currently trailing in the electoral-college swing-state polling).

1. Mark R. says:

My thoughts exactly. And he’s polling poorly in the three swing states that won him the election: PA, MI and WI. And the MI governor Whitmer’s (D) approval rating is 15 points higher than Trump’s in the state; and for some reason he thinks it’s a good idea to criticize her and stir up his cult to “liberate” the state. And there will surely be republican efforts to hurt voter turn-out, but they sure didn’t work in Wisconsin’s most recent election. Not to mention 2018 and every election held since- Republicans are losing everywhere. Trump is a huge drag on the party and his handling of this pandemic is only getting worse. He’ll still have his cult, but his cult can’t win him the election.
What will surely get Biden the Presidency is if more swing states are successful at creating a vote-by-mail system. Won’t happen in states w/ (R) governors, but I can see it happening in states with (D) governors, and that includes a number of swing states.

1. rickflick says:

Great spot. What mountains are these? I’d suggest running the power lines underground.

1. Top is Mt. Adams (~12,300 ft.). Bottom is Mt. Hood (~11,100 ft.). Photos are taken from the approximate location of our eventual retirement house (to be constructed).

The power line is running along the county road, so no luck there. Maybe (maybe) we won’t see it from our second floor.

Our neighbor wanted to run the short (300 feet, maybe?) “swing” of power line that connects to their house to provide them with power, underground rather than above ground. This was about 20 years ago:

The cost above ground: \$1000
The cost below ground: \$10,000 (20 years ago)

So, most likely, when we need a new swing to our eventual retirement house, it will be above ground. However, from either of the nearest poles, it wouldn’t obstruct either mountain view. 🙂

2. rickflick says:

Maybe you could rent a backhoe and dig the trench yourself. Get some 4″ PVC from Home Depot and ask the power company to cut you some slack. 😎

1. We have it in CA also though I believe it isn’t no-fault. It’s still my fault if I vote for the wrong person. 😉

1. Derek Freyberg says:

That’s true Paul, it is up to the voter to cast his/her/etc. vote correctly.
But California, as I’m sure you know, does have permanent absentee voting, no “good reason” required, just ask – and our primary this year was all vote-by-mail, with ballots mailed to all who registered but the option to walk in and cast a vote in person at any polling place in the county in which you reside.

1. Yes, I’ve been participating in vote-by-mail for a few years. The only downside with the primary was that I ended up voting for a candidate that had dropped out by election day.

2. Paul, that was me too (Klobuchar). But it was OK.

3. GBJames says:

I think the solution is to provide easily accessed ballot drop-off locations where you can deposit the ballots in the few days before Election Day. That way you don’t have to worry about the thing being lost in the mail or not arriving in time. ??

4. GBJames: Good idea!

I have written my State Rep about changes I’d like to see. MN is pretty conservative on things like this (in the sense of not very willing to change — either party often).

I was thrilled to get NF Absentee voting!

2. Mark R. says:

Nice part of the state. I live about 30 miles NE of Seattle. And you don’t have to sign up as permanent absentee here nowadays. It’s all mail-in; once you register to vote, you’ll get the ballot and voter guide in the mail. Once filled out, you can either drop the ballots at libraries or mail-them in free of charge. I know OR and CA also use this system. If all the states did mail-in, I doubt the US would ever have to suffer another republican president (and the GOP knows it).

1. Yes but note that the latest stimulus bill still doesn’t save the USPS. We may have to FedEx our ballots in November if the GOP gets its way.

1. Mark R. says:

Yeah, I’ve been reading about the USPS debacle. You can always rely on the GOP to use an unprecedented crisis to push through their nefarious agenda. Trump’s full-scale ban on immigration is another glaring example.

2. rickflick says:

I think the PO might soon raise rates drastically.

3. Yes. Glad I have lots of “forever” stamps. I hope they stay in business.

2. Cool, good to know!

Arlington? Granite Falls?

I mostly did nothing (on my free time) except mountaineering, kayaking, and 3-pin skiing when I lived in Renton (mainly) 1984-2001.

By the time I left, the traffic was so bad in Seattle that I was mostly choosing to just not do things, because the hassle factor was just too great. 🙁

But I sure had some goods times there. I love the marine climate. The clouds never bothered me. Our eventual retirement place (White Salmon, at about 2100 feet elev.) is less wet, warmer in summer, cooler in winter; but still nothing like Minnesota, where I live now (and grew up)!

2. Yes, never cracking 50% approval when the economy is the best it (had) been in 50+ years is pretty damned telling.

People are highly motivated to lance the Orange Abscess.

I am highly encouraged by the big turnouts in the Dem primaries.

I am cautiously optimistic.

Biden won every county in MI.

He did great in WI, nearly getting as many votes as an unopposed Trump.

I’m sure he’ll ring in well in home state PA.

And these states will likely decide the election , as in 2016.

3. “Prediction is very hard. Nate Silver is maybe the best political predicter alive, and he estimated a 29% chance of Trump winning just before Trump won.” (https://slatestarcodex.com/2020/04/14/a-failure-but-not-of-prediction/)

I really hope Trump does not win, but realistically I also hope corona numbers get better, and if they do, I fear that will make Trump look good for the short attention span that Americans predictably display.

1. Ken Kukec says:

I think you’re overestimating Trump’s chances in 2020 because you’ve fallen prey (as many of us who oppose Trump have, Kevin) to the “availability heuristic” cognitive bias, based on the outcome of the 2016 election, which was actually an outlier.

1. I suspect it might actually be the “fear heuristic”. We’ve learned from 2016 not to count our ducks before they’ve quacked. Who knows what diabolical schemes Trump and his GOP hench-people will come up with between now and November. I’m sure we’ll be wishing it was only voter suppression before this thing is over.

1. Yes, I heard about the report. I must say that I am a bit surprised that the GOP senators didn’t fill it with Trump-friendly junk. It’s a sign of the times when we ask, “What the hell happened?”, when they do something right.

6. Randall Schenck says:

Looks pretty good to me. We are all in this together, right? Just kidding, we haven’t been in this together for a very long time. Just like that other tired motto, there is more things alike than separate us. No, there is nothing alike.

7. On the pandemic curve, I find it interesting that so many seemed to have assumed that the curve would be at least roughly symmetrical about its peak; that the fall would be as fast as the rise. No reason that would be the case, of course. Even though many “smart people on TV” are warning others to not make that mistake, I suspect many will still be surprised when this pandemic goes on and on.

1. ThyroidPlanet says:

There’s also no one curve to rule them all.

It’s data. Plotted. There are shapes. (I know you understand this but I find it interesting to think about).

There’s a combination of rates into one plot, so there’s a rate of testing, there’s s rate at which the results come out, there’s a rate at which people come in for testing, etc. while what we’d like to see is just the rate of the virus hopping from one person to the next. That is impossible to measure directly.

8. rickflick says:

So, every year there are 8 or 9 ducklings fleging? At some other Botany Pond around Chicago, someone is asking, where are all these extra ducks coming from?

9. phoffman56 says:

I had predicted here that by Aug. 1, the deaths per million in US, Italy and Spain would reach around 500. It seems pretty clear that for the two mediterranean countries it will be noticeably larger. Right now for U.S. I’d still say about 50% probability that it’s between 450 and 550. The number is around 165 to 175 thousand deaths at 500 deaths per million.

I think I’d predicted about 135 per million for Canada. Probably more, say 165 per million to do 50%, say between 150 and 180.
That’s around 6,000 deaths.

I’d vastly under-predicted for Britain IIRC, but don’t remember the number. It will very likely also be over 500 per million.

10. EdwardM says:

What is a wild mallard’s average surviving brood size? I know most don’t make it through the first year, so how many on average make it to fledging, which is where I presume Dr PCC(e) loses track of them?

1. My sample size of one where my wife and I took care of 11 eggs after the mother was killed by, something, (raccoon maybe) in my backyard was 100% fledged. And my two kids made it out of the house too so I’m a heck of a two species parent.

ducks.org has an article on duckling survival rate which says less than 10% to more than 70% survive. Big factors are predation, adverse weather conditions, starvation, disease, and parasites. Apparently, lots of creatures from frog to fish to birds like to snack on ducklings. It’s a jungle out there. Interesting article at the link.

https://tinyurl.com/y9w5dd7e

11. JezGrove says:

The South Koreans are denying that Kim is seriously ill. Not sure how they usually deal with rumours about their dodgy neighbour to the north, China has issued a similar denial. Though the BBC note that neither has specifically denied that Kim has had heart surgery. https://www.bbc.co.uk/news/world-asia-52364055

12. A C Harper says:

I’ve read that Trump didn’t ask for his name to appear on the cheques… although I don’t suppose he declined.

I suspect it is more to do with bureaucrats slugging it out over ‘federal money’ and ‘state responsibilities’.

Still lots of poorer people receiving ‘free money’ from the President? Bound to have some effect.

1. Saul Sorrell-Till says:

“I suspect it is more to do with bureaucrats slugging it out over ‘federal money’ and ‘state responsibilities’.”

Why on earth would that mean Trump’s signature ending up on cheques?

2. GBJames says:

tRump said this. And if you believe it then perhaps I can interest you in this fine bridge I’ve got for sale.

1. Saul Sorrell-Till says:

“We’re gonna build a big beautiful bridge, from here to China. And the Chinesians will pay for it.”

3. I can imagine that it started when someone asked Trump, “Will you be signing the checks?”, knowing his narcissistic personality. Regardless of his answer, this plants the idea in his and his flunkies’ heads. From this point on, Trump doesn’t have to give an order to put his name on the checks. He can simply acquiesce to it. He can claim that he doesn’t know about and didn’t ask for it. That may or may not be a lie. It doesn’t really matter to Trump. I suspect something similar happened with the hiding of the destroyer, USS John S. McCain.

1. Ken Kukec says:

You really think that any flunky who came up with the plan wouldn’t immediately rush to claim credit from the Great Man for the deed, like laying a rodent at its master’s feet?

1. Sure, they might want to place the rodent, as you say, but they know that the Great Man needs to maintain some level of plausible deniability so they suck it up. A truly thankless task.

13. Simon Hayward says:

OK, hold, sell (I think your % is too high) and buy (I think it’s too low)

Hold (no data)

Sell (I doubt things will open up by then)

Sell (ditto just a different region)

Sell – but depends on the definition – just teams playing prob buy, with crowds – sell

Sell

Buy – per comment above, depends on the kids and the school system. Unmotivated kids won’t learn and motivating them, when (as in IL) their grade cannot go down is problematic. Some will learn a lot though.

TV ads – can’t comment for you – for me it’s closer to 0% as I’m unlikely to see a paid ad at all.

Honey’s chick count, since you know the egg count and will be able to assess unhatched eggs is at least a knowable outcome in the short term. I wish her well. I guess the real question is how many of her chicks will end up raising chicks of their own. Answer to that is likely close to zero in any given year. (Here endeth the cheer!)

1. Once a duck has reached fledging stage and migrates, its probability of reproducing is much higher than when you calculate that probability at birth. I am determined to get every duckling through to fledging.

I don’t know why you say the probability of her (adult) offspring reproducing is “close to zero”.

1. Simon Hayward says:

The maths (as you know) are brutal. In a stable population every adult breeding female duck is leaving two progeny. So while we may see Honey’s genes come to dominate the population it seems unlikely that most of these ducklings will live to raise their own. Clearly there is drop off at every stage, egg to hatchling to fledgling to adult to reproducing adult. But if she raises say 8 migrating offspring a year for four years that’s only about a 7% (2/32)chance on average that any one of them will do the same, (numbers obviously worse for hatchlings and eggs). Whether 7% is “close to zero” is a fair question.

2. Simon Hayward says:

In more cheery news, the Trump “rally round the flag” effect, such as it was has dissipated possibly as a result of his daily rants, his polls are back to where they were at the end of jan (about -8.8 at the moment) and sinking (for now).

14. Saul Sorrell-Till says:

Probability that I will ever truly understand probabilities: 111%

1. phoffman56 says:

I realize that’s mainly a joke.

But it’s also true, and I’m not being insulting since you are simply one of every human who ever lived, and it’s impossible to insult them all. There is at present no such thing as a decent definition of the word ‘probability’ IMHO.

If the frequentist one is someone’s counterexample, I’d say: tell me all about a specific human who actually did flip a coin infinitely often; and then I’ll sell you the Brooklyn Bridge.

If you claim the Bayesian is it, then being based on human betting, it appears that either quantum theory is dead wrong, or else it only started applying to physical reality once there were actually some humans in existence. So for something like 99.999% of time since our ‘visible’ universe came into existence (Well, the number is roughly gotten by subtracting from 1 the ratio of 300,000 over 13.6 billion, and making it into a %), quantum theory could not have applied to that physical universe.

Maybe I exaggerate: you’d need to start believing in the Everett universe as explained by David Wallace, where probability makes better sense, not worse as is often claimed.

1. There’s also the “fact” (or idea) that the human brain estimates probabilities in a manner not consistent with those laws of probability. Evolution must have looked at those laws and found them lacking.

2. Saul Sorrell-Till says:

Yes, Sean Carroll talks about what probability means in an Everettian sense in his latest book. And David Deutsch has touched on it too.

“There is at present no such thing as a decent definition of the word ‘probability’”

That’s true of practically every single concept humans have if you really look at them properly. Choice. Causation. Time. Creation. Life.

I only started getting interesting in science and philosophy in my mid-twenties, and over the last decade it’s been quite shocking to see just how many everyday concepts make absolutely no sense as they’re commonly used. So much of what we talk about is built out of sand.

1. Yes, it’s all shaped by the mechanisms evolution has planted in our heads. We make our own reality out of the real one, causation being the biggest of them all. That’s the sand on which we have built.

2. Ken Kukec says:

That’s true of practically every single concept humans have if you really look at them properly. Choice. Causation. Time. Creation. Life.

Comes a point, I think, where excogitating on such matters becomes paralyzing. And you can risk running into yourself coming around the corner. 🙂

1. phoffman56 says:

It’s obvious that not everything can be ‘defined’, since the words used would themselves need to be, giving infinite regression, or giving your “running into yourself coming around the corner”.

One needs to get started with taking, say ‘time’ and a few other things, as undefinably primitive. But not forever, if e.g. a good theory with time as emergent itself emerges someday.

To me a good example of something pretty basic which got a proper scientific definition about 100 years ago is ‘energy’.
That’s due to Emmy Noether. It’s the conserved quantity corresponding to invariance of fundamental law under time translation, and depends on the theory taking a certain common general form. Frank Wilczek in his terrific book from about 5 years ago sort of speaks of its history being reprised in his own teenage mind (and probably many others’ too). Before he learned about Noether teaching both Einstein and Hilbert with her deep theorem, it seemed to him that in the 1700s and 1800s there kept being a fudge there, with some new form of energy invented each time conservation of energy seemed to be violated. But not once Noether explained how it should be defined.

“Choice. Causation. Time. Creation. Life.”

Except for ‘time’ and ‘life’ in this list, the others seem to me to need no strict scientific definitions. And earthbound ‘life’ seems to have (modulo some choices that don’t involve ignorance, e.g. do we regard a virus particle as alive?) as being pretty well defined chemically now, in enormous contrast to 200 years ago. The ‘code’ or whatever leading from DNA/RNA towards proteins is universal, and of course is basic in the belief in a common ancestor, since there are a huge number of possibilities but only one that occurs.

But it was ‘probability’, and that somehow to me just doesn’t seem primitive in the same way that ‘time’ seems to be. But no satisfactory definition, to me anyway, other than maybe in the thoughts of Everettians. David Wallace’s book needs much more hard thinking, for me anyway. It is certainly far closer to Bayesian than to frequentist. But in practice, the latter is how almost all scientists use probability/statistics, I think.

15. Curtis says:

1. I find it unlikely Kim will die. Less than 20%. He is in his 30s and will get good medical care in China if not Korea.
2. Trump winning 75%. Elitist scorn of the deplorable will get him the midwest again.
3. MLB without fans (most likely in Arizona only) 90%. Money, money, money vs. a low risk to the players who are young and in good shape.
4. Schools absolutely need to re-open but the health risk is large. Some states (perhaps Montana, Dakotas) will open 80%. Most states 50%.

1. phoffman56 says:

But (obese in 30s)=(marathoner in 80s) maybe.

2. 2: Doubt it.

Look at the MI and WI results.

We got the Orange Abomination by just 77,000 votes distributed in WI, MI, and PA. Think he can pull that rabbit out of the hat again?

Doubt it. I can’t imagine any independent voting for Trump this time around.

16. eric says:

Probability that an average American can correctly draw a “regular” (no intervention) versus a “flattened” pandemic curve, with the axes properly labeled: 20%

I set up a toy example of flattening the curve for my 9-year-old. Had him plot out 5-6 points for each curve on a graph, then answer some simple questions about them (stuff like “if hospitals can treat 100 new cases per week, how many new cases go untreated in the first example? In the second?”). He got the point.

I’m not a teacher. It’s not hard stuff, if you’re willing to put a bit of mental effort into it.

17. Hempenstein says:

I think students will learn more. In short, they’ll be able to re-run parts of the video that seem important, and will only need to note the times of what sounds important vs. having to try to take notes on what they just heard while listening to what’s next.

Plus, the factor of being distracted by someone you’d like to sleep with won’t be there.

18. Steve Pollard says:

Kim: probably no more than 25%, Reasons: a) he has been absent without explanation before (probably gout, that time); b) if he had really been at death’s door, there would have been some effort to start preparing the ground at the parade he failed to turn up to. I reckon he’ll be getting the best medical attention his slave state can afford.

Remote learning: it is nowhere near ideal, but the outcomes for existing courses suggest that 5% is too low. I would hazard 60%

19. Probability that I will see an ad on television today telling us that “we’re all in this together” (as if that helps): 100%

For me? 0%. I don’t watch TV. I barely listen to the radio or look at news on the internet.

Haven’t watched TV, more or less at all, since 1987. (Miami Vice …)

20. Ken Kukec says:

“I came to the conclusion long ago that all life is six-to-five against.”

— Damon Runyon, A Nice Price

1. JezGrove says:

As I’ve mentioned here at WEIT in the past, I love Damon Runyon. My English teachers who had to read my creative efforts written in the present continuous, probably not so much.

1. Ken Kukec says:

I’m a Runyon fan, too. Hell, I once dated a Lit major many years ago who described me to her friend as “Runyonesque.” (Or maybe she said I had delusions of being “Runyonesque.” Who can even remember?) 🙂