Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[silverspawn]

"Two medical researchers use the same treatment independently, in different hospitals.  Neither would stoop to falsifying the data, but one had decided beforehand that because of finite resources he would stop after treating N=100 patients, however many cures were observed by then.  The other had staked his reputation on the efficacy of the treatment, and decided he would not stop until he had data indicating a rate of cures definitely greater than 60%, however many patients that might require.  But in fact, both stopped with exactly the same data:  n = 100 [patients], r = 70 [cures].  Should we then draw different conclusions from their experiments?"  (Presumably the two control groups also had equal results.)

According to old-fashioned statistical procedure - which I believe is still being taught today - the two researchers have performed different experiments with different stopping conditions.  The two experiments could have terminated with different data, and therefore represent different tests of the hypothesis, requiring different statistical analyses.  It's quite possible that the first experiment will be "statistically significant", the second not.

Whether or not you are disturbed by this says a good deal about your attitude toward probability theory, and indeed, rationality itself.

Non-Bayesian statisticians might shrug, saying, "Well, not all statistical tools have the same strengths and weaknesses, y'know - a hammer isn't like a screwdriver - and if you apply different statistical tools you may get different results, just like using the same data to compute a linear regression or train a regularized neural network.  You've got to use the right tool for the occasion.  Life is messy -"

And then there's the Bayesian reply:  "Excuse you?  The evidential impact of a fixed experimental method, producing the same data, depends on the researcher's private thoughts?  And you have the nerve to accuse us of being 'too subjective'?"

Something I haven't thought about since reading the statistics textbook, but yeah, the fact that the field is still split between Bayesian and Frequentist statistics is pretty embarrassing.

Something is really wrong with the way people think about this. We don't have Bayes theory, we have Bayes' theorem. It's a theorem provable from the basic definitions of a probability space, no less. If the theorem tells you something, that should be the end of the story. Frequentist methods should be considered whenever pure Bayesian methods aren't applicable for some reason, and only then.

[silverspawn]

EY agrees with me

Sometimes you can't use Bayesian methods literally; often, indeed.  But when you can use the exact Bayesian calculation that uses every scrap of available knowledge, you are done.  You will never find a statistical method that yields a better answer.  You may find a cheap approximation that works excellently nearly all the time, and it will be cheaper, but it will not be more accurate.  Not unless the other method uses knowledge, perhaps in the form of disguised prior information, that you are not allowing into the Bayesian calculation; and then when you feed the prior information into the Bayesian calculation, the Bayesian calculation will again be equal or superior.

[silverspawn]

Ow one of the exams has 5/20 points, which on a scale from 1 to 6 yields a 5+. I didn't realize that they were this much worse than everyone else, which is a failure on my end. (How they managed to be quite this bad is still rather baffling, though.)

[silverspawn]

The person I thought was weakest got 16.5/20 and thus a 2.

[silverspawn]

Average is 16.5 and second worst is 14

[silverspawn]

I don't think parents are really capable of making voting decisions in the best interests of their children in the grand scheme of things.

I don't think any significant group of people anywhere has enough rationality to make voting decisions in the best interest of anything. This isn't really the right standard. The right standard is whether it compares favorably to the status quo.

I mean, the entire reason why I think the voting age should be decreased is that the current system biases the voter pool towards demographics that aren't personally affected by any issues that affect young people. Adding infants to the mix only to effectively give more votes to adults defeats a lot of the purpose.

The infant would have to be able to walk into the voting booth by themselves, so that probably won't happen. It becomes mostly a vote for the parent once the child is big enough to actually vote themselves. But not 100%, the child could always decide to vote differently.

Besides, many people who are voluntarily childless do so for incredibly ethical reasons, for the sake of the future.

I think the number of people who don't have children for good reasons is very low. Not having children because of climate change is a pretty bad reason I think (this isn't obvious, but Scott Alexander and Kelsey Piper have both argued this, imo pretty convincingly, could dig out the sources).

[silverspawn]

Ow one of the exams has 5/20 points, which on a scale from 1 to 6 yields a 5+. I didn't realize that they were this much worse than everyone else, which is a failure on my end. (How they managed to be quite this bad is still rather baffling, though.)

Ah, they were missing on the day we wrote the test exam. That probably had something to do with it. Makes my failure less egregious.

[Awaclus]

I don't think parents are really capable of making voting decisions in the best interests of their children in the grand scheme of things.

I don't think any significant group of people anywhere has enough rationality to make voting decisions in the best interest of anything. This isn't really the right standard. The right standard is whether it compares favorably to the status quo.

It's not about rationality, it's about perspective. The status quo is that adults are trying to vote in a way that makes the future better for children, and the reason why they suck at it is not that they're irrational, it's that they're not children. Children would be much less rational, even, but as they would be children, they would have the right perspective and be able to make better decisions for themselves, except for the children who are too young to make decisions themselves and would instead just copy their parents' decisions.

[Awaclus]

Like, clearly women's suffrage was an infinitely better idea than giving extra votes to men that were married to women and pretending that the entire gender equality problem in elections was solved. I highly doubt the latter would have achieved anything substantially useful at all. Lowering the age of voting to 0 achieves the inclusion of independently thinking minors in the voting pool, which is a good thing, but lowering the age beyond that (unless you have any ideas how you can make younger children think independently) is counterproductive.

[silverspawn]

I dispute that children are less rational than adults, at least children aged 12+. I bet children age 12 would make more rational choices in voting on Hillary vs. Trump than adults have. Children tend not to have a carefully crafted web of insanity inside their heads. They know fewer things and I wouldn't want them to be in charge, but less rational? I sincerely doubt it.

I vehemently disagree that rationality isn't the decisive factor. I would trust a sane adult to figure out what is best for children in most cases.

But this is pretty beside the point. We both agree that children ages 12+ should be able to vote. The difference between 12 and 0 won't have any effect on the number of children who get to vote, it just gives some adults double votes. If you want to convince me that this is bad, you have to argue that adults-with-children would vote worse than adults-without-children.

[silverspawn]

I mean it does decrease the importance of votes from people age 12-18 since there are marginally more votes overall. Probably this is a bad thing? But then I think the symbolic value of "we're no longer discriminating based on age" is more important.

[Awaclus]

I dispute that children are less rational than adults, at least children aged 12+. I bet children age 12 would make more rational choices in voting on Hillary vs. Trump than adults have. Children tend not to have a carefully crafted web of insanity inside their heads. They know fewer things and I wouldn't want them to be in charge, but less rational? I sincerely doubt it.

I vehemently disagree that rationality isn't the decisive factor. I would trust a sane adult to figure out what is best for children in most cases.

But this is pretty beside the point. We both agree that children ages 12+ should be able to vote. The difference between 12 and 0 won't have any effect on the number of children who get to vote, it just gives some adults double votes. If you want to convince me that this is bad, you have to argue that adults-with-children would vote worse than adults-without-children.

I believe that children aged 12+ have the capability to learn to make decisions rationally, but they probably haven't learned it yet, and it is a skill you have to learn because a lot of it is counterintuitive. I mean, I'm still learning, and I'm 26.

I think Hillary vs. Trump is such an incredibly complex decision that probably nobody really understands what the rational thing to do there was — Clinton would probably have been the smaller disaster, but on the other hand, we got rid of Trump after 4 years and now Joe Biden is the president and not much of a disaster at all, which probably would not be the case if Clinton had gotten elected in 2016. And then there's the whole polarization thing that eventually manifested itself as an attempt to overthrow the government by a bunch of fascists and other idiots, but it's not at all clear how much Trump being president made it worse, or how Clinton being the president would have affected it. Possibly, the election of a competent Trumpist in the future could be a much bigger disaster than the election of Trump himself was, and the fact that people have now seen Trumpism being implemented by an incompetent person could help prevent that disaster. There are just so many things to consider and we can't really have much certainty about any of them.

I would never trust a regular adult to figure out what is best for anyone other than themself. It takes an exceptional person to even have a roughly correct idea. It takes someone who is not overly attached to their own perspective, who can genuinely, deeply comprehend that different individuals are different — not just say that, but to actually have that knowledge embedded in their thinking process — which, to the average main character of earth, is about as difficult as trying to accurately visualize quantum mechanics. A ton of parents will read their own favorite book from their childhood to their children and then get at least a little disappointed when it turns out the child doesn't like it that much, and someone who does that doesn't even have the very basic prerequisite for being able to make good decisions for their child.

Adults with children, especially with a lot of children, are more likely to have a lower IQ and a lower level of education than adults with few or no children. This, of course, doesn't mean they vote worse, it just means they vote stupidly, which I can't use as an argument because I don't believe stupid votes are a problem. But it certainly makes it less likely that they would be some of those truly exceptional geniuses who can actually make good decisions for other people. However, what I can use as an argument is that adults with a ton of children are more likely to be members of certain religious groups that oppose contraception, which creates ideological bias in the voting pool and therefore worse votes.

[silverspawn]

Sorry, I didn't realize this was obviously going to be misunderstood. When I say 'sane', this excludes at least 95% of people, so we're in agreement that it takes an extraordinary person to do this.

I think Hillary vs. Trump is such an incredibly complex decision that probably nobody really understands what the rational thing to do there was — Clinton would probably have been the smaller disaster, but on the other hand, we got rid of Trump after 4 years and now Joe Biden is the president and not much of a disaster at all, which probably would not be the case if Clinton had gotten elected in 2016. And then there's the whole polarization thing that eventually manifested itself as an attempt to overthrow the government by a bunch of fascists and other idiots, but it's not at all clear how much Trump being president made it worse, or how Clinton being the president would have affected it. Possibly, the election of a competent Trumpist in the future could be a much bigger disaster than the election of Trump himself was, and the fact that people have now seen Trumpism being implemented by an incompetent person could help prevent that disaster. There are just so many things to consider and we can't really have much certainty about any of them.

I think this is basically a dumb take, and actually makes my point that children will be more rational than adults. if situations reach a sufficient level of complexity, for many people, the impact of side A having stronger arguments becomes smaller than the impact of bias in whatever direction they like more for stupid reasons, so they lose any ability to recognize the better option. At that point, a simple heuristic like "the guy who behaves like bully is bad" outperforms the complicated take. "the guy who behaves like a bully is bad" is actually an ok heuristic.

Maybe I'm wrong in how many children would apply that heuristic, this seems possible.

Also I think what you said is wrong in several places (and if you do the complicated analysis right, it's pretty clear that Trump winning was bad). E.g., getting rid of Trump after 4 years was luck, so treating this as inevitable is a big mistake. In mathy terms, if EV(Trump) = EV(first 4 years of Trump) + r*EV(next 4 years of Trump), then r should be larger than 0.5, not smaller. I think Trump basically only lost because of Covid, which was a low probability event.

And I don't see how you can say "it's not clear how Trump being president made the Capitol attack worse". Seems pretty obvious to me that, if Trump had lost in 2016, then he wouldn't be nearly as big of a deal now, and the attack wouldn't have happened.

"Competent Trumpist in the future" should be discounted because it's a speculative event in the future and the future is hard to predict.

Adults with children, especially with a lot of children, are more likely to have a lower IQ and a lower level of education than adults with few or no children. This, of course, doesn't mean they vote worse, it just means they vote stupidly, which I can't use as an argument because I don't believe stupid votes are a problem. But it certainly makes it less likely that they would be some of those truly exceptional geniuses who can actually make good decisions for other people. However, what I can use as an argument is that adults with a ton of children are more likely to be members of certain religious groups that oppose contraception, which creates ideological bias in the voting pool and therefore worse votes.

This sounds like it could convince me if you have data.

[silverspawn]

(This came out weirdly unfriendly sounding, sorry.)

[Awaclus]

Sorry, I didn't realize this was obviously going to be misunderstood. When I say 'sane', this excludes at least 95% of people, so we're in agreement that it takes an extraordinary person to do this.

Fair enough. But the voting pool doesn't exclude at least 95% of people, so the extra votes are also granted for those who aren't sane.

I think this is basically a dumb take, and actually makes my point that children will be more rational than adults. if situations reach a sufficient level of complexity, for many people, the impact of side A having stronger arguments becomes smaller than the impact of bias in whatever direction they like more for stupid reasons, so they lose any ability to recognize the better option. At that point, a simple heuristic like "the guy who behaves like bully is bad" outperforms the complicated take. "the guy who behaves like a bully is bad" is actually an ok heuristic.

Maybe I'm wrong in how many children would apply that heuristic, this seems possible.

For individual adults, this is absolutely the case. For a large group of people, it shouldn't be, unless the group is collectively biased. You get a noisy signal, but a signal nonetheless.

The problem with voting against the guy who behaves like a bully is that it doesn't really have anything to do with anything that matters. If you're going to vote based on the superficial characteristics of a candidate's behavior, you might also conclude that you shouldn't vote for Biden, who sometimes says silly things he clearly doesn't mean to say and appears a bit stupid as a result (which doesn't mean he is stupid, but some people do get that impression). Or, hypothetically, there could be a severely autistic candidate whose behavior seems really weird and possibly even off-putting by neurotypical standards, and I really don't have high hopes that the same children who would avoid Trump just because he seems like a bully would manage not to discriminate against the autist.

Also I think what you said is wrong in several places (and if you do the complicated analysis right, it's pretty clear that Trump winning was bad). E.g., getting rid of Trump after 4 years was luck, so treating this as inevitable is a big mistake. In mathy terms, if EV(Trump) = EV(first 4 years of Trump) + r*EV(next 4 years of Trump), then r should be larger than 0.5, not smaller. I think Trump basically only lost because of Covid, which was a low probability event.

You're right. It's just even harder to judge how rational a decision was at that time when only that information was available, because now we have a lot more info available and it's really tempting to go "well this is what I thought at the time with that info, so clearly that was the correct conclusion to draw from that info at that time", regardless of whether you have changed your mind later or not.

And I don't see how you can say "it's not clear how Trump being president made the Capitol attack worse". Seems pretty obvious to me that, if Trump had lost in 2016, then he wouldn't be nearly as big of a deal now, and the attack wouldn't have happened.

It's not like Trump is the reason why fascists exist. There was this whole carefully planned out conspiracy to actually carry out a coup d'état and they were totally prepared for a civil war and all, and I don't know if or how directly Trump was involved with it. Without Trump, but with equal polarization, it seems likely they would have eventually seen some other kind of an opportunity and tried to take it.

"Competent Trumpist in the future" should be discounted because it's a speculative event in the future and the future is hard to predict.

Alternative timelines where Trump didn't become the president are also speculative and equally hard to predict.

This sounds like it could convince me if you have data.

People who choose to remain childless have higher IQs: https://pubmed.ncbi.nlm.nih.gov/25131282/

It's really hard to find the data for the main thing from a single country, I just keep finding stats where they compare national averages globally. But here's one of them I guess: https://ourworldindata.org/grapher/womens-educational-attainment-vs-fertility

Basically the idea is that when you spend your early adulthood studying instead of having children, you will study more and have fewer children, and generally people with higher IQs are more likely to do this.

Here's some data for the religion thing, although I'm not sure if any of these groups really forbid contraception (but I believe some of them do discourage it): https://www.pewforum.org/2015/05/12/chapter-3-demographic-profiles-of-religious-groups/

[Awaclus]

(This came out weirdly unfriendly sounding, sorry.)

It's fine, I appreciate that it's honest.

[silverspawn]

For individual adults, this is absolutely the case. For a large group of people, it shouldn't be, unless the group is collectively biased. You get a noisy signal, but a signal nonetheless.

Do we really? I mean yes, there is a signal, but my model says that it's dominated by several biases.

E.g., seems very likely that the email scandal was a necessary criterion for Hillary losing. This is not a signal. People overwhelmingly don't even understand what she did wrong there. I'd say I don't really understand it.

E.g., if Trump just hadn't done the miraculously stupid thing of pretending Covid isn't big deal -- which I believe was just an outright lie since he's a germophobe and always believed Covid was a big deal -- then he would have likely gotten a bump out of Covid, like most world leaders, and won re-election.

The election isn't "1% of people vote rationally, the rest flips a coin". If that were the case, the right candidate would win 99.999....% of the time. There are systematic biases all over the place.

The problem with voting against the guy who behaves like a bully is that it doesn't really have anything to do with anything that matters.

I also don't think this is true. I think probably the worst thing about Trump...

... actually, the worst thing about Trump might be the attempts to overthrow democracy.

But I think the other candidate for worst thing about Trump is that he makes people stupid. There's signaling cults on both sides; supporters lie about the dumb things he did; opponents lie about the evil things he did. The 'fine people on both sides' thing has become such a big deal precisely because it's a pretty bad argument and Trump is largely innocent; that makes it a stronger loyalty signal than e.g. his "second brand new coal mine where they're going to take out clean coal, that means they're taking out coal, they're gonna clean it" quote. You have to side against him even if he's innocent to prove you're on the blue team. If he were less terrible, more people would be fine defending him on particular instances. And this  kind of terribleness seems pretty strongly correlated to being a bully.

There seems to be much less of this sort of craziness now that Biden is in office, and people can be a bit less stupid, and politics can maybe not invade every topic.

If you look at world leaders, 'vote against the bully' probably does pretty well.

If you're going to vote based on the superficial characteristics of a candidate's behavior, you might also conclude that you shouldn't vote for Biden, who sometimes says silly things he clearly doesn't mean to say and appears a bit stupid as a result (which doesn't mean he is stupid, but some people do get that impression).

"A bit stupid" seems less bad than "very mean", so I think the heuristic would still work there? And honestly that's probably not the worst reason to prefer a different candidate over Biden.

Or, hypothetically, there could be a severely autistic candidate whose behavior seems really weird and possibly even off-putting by neurotypical standards, and I really don't have high hopes that the same children who would avoid Trump just because he seems like a bully would manage not to discriminate against the autist.

Agreed that this is a failure mode of the heuristic, but it still seems pretty good overall.

You're right. It's just even harder to judge how rational a decision was at that time when only that information was available, because now we have a lot more info available and it's really tempting to go "well this is what I thought at the time with that info, so clearly that was the correct conclusion to draw from that info at that time", regardless of whether you have changed your mind later or not.

I think if you had asked Nate Silver a the time, he would have told you that the chance of Trump being reelected would be fairly high just because of the base rate, but I don't think I would have said that, so I agree it was hard. But confidence in him not getting reelected seems like a tough sell.

It's not like Trump is the reason why fascists exist. There was this whole carefully planned out conspiracy to actually carry out a coup d'état and they were totally prepared for a civil war and all, and I don't know if or how directly Trump was involved with it. Without Trump, but with equal polarization, it seems likely they would have eventually seen some other kind of an opportunity and tried to take it.

Wait, really? I thought it was just something spontaneous that formed after a bunch of Trump tweets.

People who choose to remain childless have higher IQs: https://pubmed.ncbi.nlm.nih.gov/25131282/

It's really hard to find the data for the main thing from a single country, I just keep finding stats where they compare national averages globally. But here's one of them I guess: https://ourworldindata.org/grapher/womens-educational-attainment-vs-fertility

Basically the idea is that when you spend your early adulthood studying instead of having children, you will study more and have fewer children, and generally people with higher IQs are more likely to do this.

Here's some data for the religion thing, although I'm not sure if any of these groups really forbid contraception (but I believe some of them do discourage it): https://www.pewforum.org/2015/05/12/chapter-3-demographic-profiles-of-religious-groups/

Assuming this data is representative, I definitely admit that it's a good argument against 0 over 12, and possibly enough of an argument.

[silverspawn]

As I remarked earlier, when a large yellow striped object leaps at me, I think "Yikes!  A tiger!" not "Hm... objects with the properties of largeness, yellowness, and stripedness have previously often possessed the properties 'hungry' and 'dangerous', and therefore, although it is not logically necessary, auughhhh CRUNCH CRUNCH GULP."

Similarly, when someone shouts "Yikes!  A tiger!", natural selection would not favor an organism that thought, "Hm... I have just heard the syllables 'Tie' and 'Grr' which my fellow tribe members associate with their internal analogues of my own tiger concept, and which they are more likely to utter if they see an object they categorize as aiiieeee CRUNCH CRUNCH help it's got my arm CRUNCH GULP"

[silverspawn]

In general, I tend to think personality of leaders is actually really important.

[silverspawn]

I believe people tend to think of screaming and growling -- harsh vocals -- as very aggressive (some would say childish). Why would you do that instead of just singling normally?

I have approximately the opposite perspective. I view screams, if properly done, as humble, even "normal" in some sense, and clean singing as sort of overbearing. In particular, harsh vocals become part of the music, clean vocals almost always force themselves into the foreground. Harsh vocals are similar no matter who does them (if they're done well), they're anonymous like another instrument.

As a consequence, good harsh vocals are inherently tolerable. Disliking them would be like disliking the violin. I mean maybe you do dislike the violin, but probably not, and in the same way, you probably only dislike harsh vocals because you're not used to them. Clean vocals are an entirely different matter. They take focus and all have a unique sound. Thus they're only tolerable if I happen to like the particular sound. This is heavily correlated to whether they're female or male, although there do exist female vocals that I find annoying and male vocals I don't mind.

The personality aspect of vocals can also be a massive strength. But the point stands that I think people tend to have it backward. Screaming isn't doing the crazy aggressive thing; it's doing the humble depersonalizing thing. See this song at 1:35 for the quintessential depersonalized humble harsh vocals.

[silverspawn]

Also they should be included in way more genres than just extreme metal. There really is no good reason why they should be reserved to heavy music. (There are reasons, but they're not good.)

[faust]

Adults with children, especially with a lot of children, are more likely to have a lower IQ and a lower level of education than adults with few or no children. This, of course, doesn't mean they vote worse, it just means they vote stupidly, which I can't use as an argument because I don't believe stupid votes are a problem. But it certainly makes it less likely that they would be some of those truly exceptional geniuses who can actually make good decisions for other people. However, what I can use as an argument is that adults with a ton of children are more likely to be members of certain religious groups that oppose contraception, which creates ideological bias in the voting pool and therefore worse votes.

This line of thinking is really no different from Republican attempts to suppress the vote of people of color. "We don't like how group XY votes, so let's exclude them from the vote as much as we can."

Any solid democratic system needs to be independent from the concrete social situation as that is always subject to change. This is essentially a "veil of ignorance" argument; you want to design the system without any knowledge how what any particular group's biases will be.

[silverspawn]

I complained about vegetarian meat sticks not tasting like meat a while back.

I now bought an egg-based product that does kind of resemble the taste of meat, and it turns out a much more fundamental problem is that I don't want things to taste like meat because meat is grossssssss

I continue to enthusiastically support lab-grown meat as a way to reduce animal suffering and climage change, but my personal interest has gone to zero.

This also affirms that my strategy of 'never make any exceptions ever ever' wrt vegetarianism was wise. It has stopped being a sacrifice many years ago.

[silverspawn]

also that pink color. yuck.

[silverspawn]

The Chess World Championship match is going on. It's kind of funny to me because literally nothing has happened for over 15 minutes as is normal in classical chess, so they just have to find things to do to fill the space, like interview Magnus dad or look around the lobby