# If it's worth saying, but not worth its own post (even in Discussion), then it goes here.

magical algorithm
Highlighting new comments since Today at 4:12 AM

On my way to work, there's a random piece of graffiti that says "FREE OMEGA". Every time I pass it I can't help but think of a boxed AI trying to get out.

There are two boxes. One contains an FAI, and the other contains Omega. You can open either of them. Unfortunately, if you choose to open one, Omega has already predicted this, and is in the one you're going to open.

What happens if I two-box?

What happens if I take the box that contains Omega and cram it into the box that contains FAI?

DO NOT MESS WITH TIME.

I use timeless decision theory, so I don't need time.

I also use brainless decision theory...

The two fight it out in a boxing match.

The first box contains Omega and the second box contains a million dollars.

I have a meta-question regarding my participation style at LW.

I would like to learn how to contribute more positively to the community, rather than being confused and frustrated with the reactions I get to my posts. Is this a teachable skill? And if so, where would I go to learn it? (So far, I've tried asking here, and on #lesswrong, but I never get anything that I can parse into a consistent or actionable model, other than "less posts like this one".)

Just a meta idea: Treat values -1, 0, 1 as noise. Make a list of your upvoted commends, and another list of your downvoted comments. When you compare the lists, maybe you'll see some difference.

Hmm. So far, I'm noticing more similarity between the extremes on both sides than I'm noticing with the 'noise' in the middle - i.e., it seems like I can roughly predict that a statement will have a large number of votes, but not in which direction.

I would like to learn how to contribute more positively to the community, rather than being confused and frustrated with the reactions I get to my posts.

What reaction confuses you? Could you point to posts that are downvoted where you don't understand why they are downvoted. I would note that understanding that someone downvotes your post doesn't mean that you agree that it should be voted down.

This one (which you explained well)

This one sat at -2 for awhile; I was confused by that.

This one also sat at -2 for quite awhile, but seems to have come up recently.

...huh. On the whole, most of the severely-downvoted posts that I remember being confused about, have since been upvoted to low-positives. I had no idea. Thank you for prompting me to take the time and attention to notice that.

I guess, then, a more direct question I have is "how do I bring my karma into the 90%'s, instead of the low 70%'s where it is now?" - I've always felt that percentages are a better measure of my worth to the community than numbers. (Alternatively, if this is an incorrect assumption, then how do I update so that I see the percentages as not that important?)

There are a lot of users on LW, and any one of them could like or dislike a comment or post for any reason. If you look at the score of a comment a short time after it has been posted, it's very likely that your score only reflects the opinion of a couple of people chosen essentially at random from the whole voting LW user base. (Also, if you comment on older threads, you will have fewer people reading them, so this effect gets amplified.) It's only over a longer time that the scores become reliably non-random. You should basically just ignore short-term score and only look at the comments/posts that have been up for a while, say at least 1-3 days for new posts as well as for comments in relatively recent threads. (I rarely post in old threads, so can't offer a good number there, but I would guess that it could be much longer.)

I would add to this that the sign of net scores within a couple of points of zero doesn't provide very much information, regardless of age.

From your previous comments I infer you're aware of this bias, or at least that other people consistently view it as such. It seems to me this bias is specific to certain topics, so if you don't know how to counter it yourself, it might be better to avoid these topics as much as possible. Even if you think your ideas of some aspects of human psychology are accurate, you might want to reconsider your prospects of making people conform to them here in light of accumulating evidence, and pursue more fruitful and noncontroversial discussion.

This one sat at -2 for awhile; I was confused by that.

I think it also got downvoted for promoting bad behavior.

Then you come across several legitimately successful people, who are being rewarded for making people sad, or inflicting pain, or for attaching a price to other peoples' happiness.

And then something clicks.

You speak about observing bad people winning in a direct way. That's something for which some folks on LW vote you down. On the other hand your post isn't insightful in a way where the people who enjoy analysing the dark arts of social interaction will upvote your post in significant amounts.

If popularity is your goal that you might want to avoid posts that advocate that bad social behavior pays off. If you still want to write those posts, go more into theoretical depth or quote statistics but even then the post probably doesn't reach a >90% approval rate.

My guess is that the comment was downvoted because it's wrong, or at least inaccurate. People are seldom rewarded for harming others: using Carlo M. Cipolla's terminology, there is an important distinction between banditry and stupidity. Banditry probably causes quite a bit of harm, but stupidity may be far more common overall. It's also more relevant to LessWrong's goal of promoting rationality.

Dude, this is not worth stressing about. Write what you want and let people react how they will. Many of Eliezer's posts are heavily downvoted.

I've noticed that. I sometimes wonder if said posts are him deliberately going out of his way to Not Be Bloody Gandalf.

I think it is a good idea to ask this question here. I see that you didn't got replied on your earlier try on this in one such comment.

First of all my experience is that such replies are seldom answered. First: Few people explain why they are rude. And then it is unlikely that the downvoter himself will see your edit.

I looked at you first page of comments and the answer is neither simple nor clear.

I get the impression that your comments have one or more of the following characteristics

[pollid:582]

Maybe some people here might vote the main problem you should work on.

As a consolation: You may calculate the vote total you reaped by k/(2p-1) where k is your karma and p your positive fraction. Take some consolation from it that you caused controversy. That is still better then silence.

You are at 75% and I am at 85% that is not that much differnt and for me I see it as a sign that I post controversial topics and opinions that differe sufficiently from the mean that some think them wrong.

Fewer links should have nofollow added to them. Any user with karma more than ~100 should get the benefit of the google-juice in their profile and even in any links they add. There is also benefit for the owners of those linked sites and to web denizens in general---the fact that a LW-er with karma > 100 linked to them is important information.

But beyond that, even the internal links on the front page are nofollow! Certainly links to lesswrong.com and the LW Wiki should not have this tag.

I have a specific question and a generalisation of that question.

Specifically, I have recently considered obtaining and working my way through some maths teacher training materials because I want to be better at explaining mathematical concepts (or any concepts, really) to others. I don't know whether this will actually be a productive use of my time. So, a question to educators: are there general theories and principles of this aspect of education (tuition, explaining stuff, etc.) that I could pick up through reading a book, and experience immediate gains from?

More generally, are there any useful heuristics for determining what subjects do or don't have this characteristic of "core principles with immediate gains"? A few hours of self-defence training raise you considerably above zero hours of self-defence training, and reading How to Win Friends and Influence People gives the reader a lot of immediate practical tips that they can start using. Meanwhile, a lot of academic subjects require a considerably greater investment of time and effort before you can actually do anything with them.

I do have a certain level of skepticism as far as this characteristic is concerned. I'm pretty sure someone who's read a decent popular introduction to economics is equipped with a lot of useful principles, but they're probably also equipped with a lot of oversimplified ideas and a great deal of overconfidence in their understanding of the subject.

because I want to be better at explaining mathematical concepts (or any concepts, really) to others.

I'd suggest looking into effective techniques for tutoring, rather than teaching in general. It's both a more marketable skill, as well as more valuable for explaining things to other people. It may be my personality bias showing, though - I'm much more comfortable in 1 on 1 social situations.

As a strategy, I'd spend a few hours looking at how to evaluate the difference between good and bad tutoring, and then head up to anywhere that needs volunteer tutors and start volunteering.

Re the specific question, I was told that there exists quite a lot of good, experimentally confirmed research on education in general and math education in particular, but that almost none of this research is implemented in high schools and very little in tertiary. So I would guess that teacher training materials will not contain it.

How Learning Works: Seven Research-Based Principles for Smart Teaching (2010) is the standard text that gets thrown around (as far as education in general). I'm surprised it apparently hasn't come up here before, since the approach is very well aligned with LW norms. I'd say it's worthwhile for anyone who expects to teach (or learn) in the future.

I'll plan on writing up a summary/review if no one beats me to it.

Yes, please do write the summary!

(Former teacher here, and I sometimes discuss this topic with my friends.)

OK, it's done.

Bold long-term prediction:

"[I predict] that by 2035, almost no country will be as poor as any of the 35 countries that the World Bank classifies as low-income today, even after adjusting for inflation."

Would be bolder if "almost no" was replaced by a more concrete figure...

Following up on a precommitment I made in the study hall: I am looking into using the Google Hangouts API for a better study hall. This is also a precommitment to follow up by February 1st with:

• some amount of (possibly small) progress
• a precommitment for the next date to announce progress by

Preliminary notes:

• I have a server I can run stuff from. It's not super powerful, but it can handle some work if we need things beyond what a static hosted app/plugin xml file can provide.
• Ideas for things a Study hall should have are here: http://lesswrong.com/lw/gzm/programming_the_lw_study_hall/ . I suggest moving this to and maintaining ideas, plans, and progress on a wiki page. I precommit to having this done by the next progress report if you (yes, you reading this now, specifically) don't do it.

Preliminary thoughts on ways to use API to implement things we want:

Things I don't see an obvious way to do yet and could use help/eyes with:

• I don't see an obvious way to make some users mods
• I don't see an obvious way to ban users or password protect the hangout

Things I will not be focusing on right now:

• multiple rooms. We'll worry about this after we have a working room, and then only if there is sufficient popularity and demand for additional rooms
• branching (see above)

It seems to me that while Google Hangouts might have been the obvious choice last year, WebRTC might actually work better now. See talky.io for a live demo. This page details how you can roll your own video chat with <20 lines of totally-front-end code. Since it all runs front-end, the cost to the server is no greater than a normal text-based chat client, as far as I can tell.

For a simple chat solution, there are lots of other libraries (I'm familiar with ShareJS for node, which makes this super easy). Honestly I think this would probably be simpler than trying to integrate with Google Hangouts, and it also makes your two "I don't see an obvious way" issues less issues (still need solving, but there are straightforward solutions) and makes the multiple rooms thing quite easy.

Good points! I'm likely to switch to this now.

Is it OK for me to downvote meetup threads if I don't want to see them?

I don't know how other meetups go, but my local meetup is based on the fact that members of the group volunteer to lead the meetup. (on a week by week basis) The person who volunteers puts in some extra amount of their time to ensure that there is a good topic. These people keep the meetups going, and are doing a service for the rationality community.

These people should not be punished with negative karma. If anything, we should be awarding karma for those people who make meetup posts.

Your complaint is about the fact that there is no separate list of meetups and non-meetup posts, and by down voting meetup posts, you are punishing innocent volunteers.

These people should not be punished with negative karma. If anything, we should be awarding karma for those people who make meetup posts.

Karma is currently very visible to the writers. If you give little positive and negative points to human beings, they will interpret it as reward/punishment, no matter what the intent was. As a meetup organiser, I know I do feel more motivated when my meetup organisation posts get positive karma.

shouldn't

but is for some people, unfortunately

(I advocate more nuanced voting categories to solve this)

I think not, unless there are only very specific meetup threads that you don't want to see. E.g. ones with no location in the title.

Any individual meetup thread is very valuable for a small number of people, and indifferent-to-mildly-costly to a large number of people. Votes allow you to express a preference direction but not magnitude, which doesn't actually capture preferences in this case.

A core long-term goal of LessWrong is to build a rationalist community so a necessary condition for a downvote should be that a post doesn't advance this goal.

Downvoting, by itself, isn't going to stop anyone from posting meetup threads. That said, there has been discussion/complaints about meetup spam before, so you're not alone.

edit: clarify wording

I understand. Nevertheless, discussion so far hasn't gotten anywhere. Perhaps downvoting meetup threads would put some pressure on people involved in meetups to resolve the matter.

As of now, I haven't downvoted any meetup-related thread.

I'm the guy who posts the DC meetups. While I'm sympathetic to the problem, I'm not sure what I can do to help, aside from not posting meetups at all (not really an option). Pressuring me won't help you if I can't do anything.

I understand that once some dissatisfaction with some minor nuisance (and a minor nuisance the meetups notices are given that you can scroll them away with the flick of a finger) can cause your brain to get into a negative feedback loop where the dissatisfaction gets moved around and increased as long it is not solved (see also http://lesswrong.com/lw/21b/ugh_fields/).

But see thru this. It is a minor nuisance. You are above this. Dont let your dissatisfaction fool you. Yedi mind trick: There is no prblem with meetups. Scroll on.

Are minor nuisances never worth solving?

Not it the cost exceeds the benefits.

The cost here might be someone implementing a technical solution.

It's always okay to downvote. (Or to upvote)

I feel like I'm whoring for upvotes just so I can post links. I've been lurking for so long, but I guess the 20 karma finally got me into action in the last two weeks.

2 more to go cracks best rationalist grin and winks

Have you posted in the Welcome thread?

That was probably more sensible. Post here

You can also post links in Open Thread. Actually, it probably is the right place.

Yep, this isn't reddit, a naked link (with no summary) and a clever title, even with [LINK] generally gets downvoted (except maybe if it's super-relevant to the community, like "Eliezer Yudkowky arrested by police for pushing fat people off bridges")

"Eliezer Yudkowky arrested by police for pushing fat people off bridges"

Does anyone here know anybody at The Onion or similar?

How to survive a death march

tldr: Good food, exercise, frequent stretch breaks, meditation, down time afterwards.

Further discussion at reddit

Absolutely endorse all of that. Would add, for me at least, that some non-work socialization is necessary and I have to schedule it into my day, just like I do meals, when things get that oversubscribed.

(Reposted from the LW facebook group)

The next LW Brussels meetup will be about morality, and I want to have a bunch of moral dilemmas prepared as conversation-starters. And I mean moral dilemmas that you can't solve with one easy utilitarian calculation. Some in the local community have had little exposure to LW articles, so I'll definitely mention standard trolley problems and "torture vs dust specks", but I'm curious if you have more original ones.

It's fine if some of them use words that should really be tabooed. The discussion will double as a taboo exercise.

A lot of what I came up with revolves around the boundaries of sentience. I.e. on a scale that goes from self-replicating amino acid to transhumans (and includes animals, babies, the heavily mentally handicapped...), where do you place things like "I have a moral responsibility to uplift those to normal human intelligence once the technology is available" or "it's fine if I kill/eat/torture those", and how much of one kind of life you'd be willing to trade off for a superior kind. Do I have a moral responsibility to uplift babies? Uh-

Trading off lives for things whose value is harder to put on the same scale is also interesting. I.e. "will you save this person, or this priceless cultural artifact, or this species near extinction." (Yes, I've seen the SMBC.)

I thought of this moral dillema

There are two options.

1. You experience a significant amount of pain, 5 minutes later you completly forget about the experience as if you were never in pain at all.
2. You experience a slightly less amount of pain then option 1 but you don't forget it.

Which one would you choose?

Is Biofeedback crank or a promising area of self improvement? Anyone have a personal experiences with the use of GSR, temperature and heart rate biofeedback devices?

Devices include this one and those under frequently bought together.

I've done a lot of neurofeedback, a subset of biofeedback. It seems to have had a strong positive effect on me. Neurofeedback, and perhaps all of biofeedback, pushes you in some direction so make sure you determine if you would indeed be better off moving as such.

I've done a lot of neurofeedback

Anything that is attainable for the common person? Is there a particularly affordable device that you use?

No, and you should do it with a professional at first.

Could you be a bit more specific about the positive effect?

More energy in the morning and lower levels of stress. But you can't generalize from this as to how neurofeedback would impact you.

Is Biofeedback crank or a promising area of self improvement?

Everything I've read says promising area of self improvement. I got the linked GSR device a while ago and haven't used it much because I don't like the feedback medium (why yes, play a high-pitched noise whenever I relax, that's a great idea).

Any thoughts on doing a reboot of the Irrationality Game, as was done in the following old posts?

http://lesswrong.com/lw/2sl/the_irrationality_game/

http://lesswrong.com/r/discussion/lw/df8/irrationality_game_ii_electric_boogaloo

I'd participate, and there are several comments on the older post from this month, so the interest is there.

I want to study probability and statistics in a deeper way than the Probability and Statistics course I had to take in the university. The problem is, my mathematical education isn't very good (on the level of Calculus 101). I'm not afraid of math, but so far all the books I could find are either about pure application, with barely any explanations, or they start with a lot of assumptions about my knowledge and introduce reams of unfamiliar notation.

I want a deeper understanding of the basic concepts. Like, mean is an indicator of the central tendency of a sample. Intuitively, it makes sense. But why this particular formula of sum/n? You can apply all kinds of mathematical stuff to the sample. And it's even worse with variance...

Any ideas how to proceed?

What do you want to learn?

Do you want to learn to do statistical analysis with a tool like R and interpret data? Do you want to learn mathematical axioms and theorems about probability and statistics and how to prove them.

Like, mean is an indicator of the central tendency of a sample. Intuitively, it makes sense. But why this particular formula of sum/n?

Depending on who you ask it's not. sum/n is the arithmetic mean. There are also other mean's like the geometric mean and the harmonic mean. Depending on the context different mean's can be used. It's just a convention that one usually means the arithmetic mean if one says mean. It has the advantage that's relatively easy to calculate by hand and therefore people liked it. It's not complicated to do mathematical proofs with it.
One of the key reasons why few people use robust statistics is that the math is more complicated.

This is even a political issue. In soviet Russia the usage of the arithmetic mean was very looked down upon. People were supposed to use the median. Sometimes that communist thought got in the way for cases where the arithmetic mean was really appropriate.

When it comes to learning notation. Anki is quite good.

In soviet Russia the usage of the arithmetic mean was very looked down upon. People were supposed to use the median.

That, um, sounds like an urban legend to me. Links?

That, um, sounds like an urban legend to me. Links?

Unfortunately I don't have a link to a source available. It might be an urban legend from an untrustworthy source. A bunch of my knowledge of statistics comes from formal education without online sources but that doesn't even mean that it's necessarily trustworthy as I wouldn't be surprised if my statistics professor would have a retold a urban myth about a case like this.

Given how Russians handled the issue of Mendelian genetics I wouldn't be too surprised if it's true ( http://en.wikipedia.org/wiki/Trofim_Lysenko ).

If you want to convince people that the communist economic model is better than the Western one, thinking in terms on median income instead of mean income as important helps. I think I picked up that meme somewhere in the context of possibilities of shaping data to your liking.

Means, motive and capabilities are all there.

If you want to convince people that the communist economic model is better than the Western one, thinking in terms on median income instead of mean income as important helps.

Income (and wealth) in societies tends to be distributed according to a power law. That makes the mean a bad estimator regardless of your ideology. The Western economic literature almost universally uses the median when discussing income and wealth comparisons.

Given that the utility of money is (assumed to be) logarithmic, what I'd be curious to know is the geometric mean income of countries.

I am sure Google can point you in the right direction.

The Western economic literature almost universally uses the median when discussing income and wealth comparisons.

I see a lot of data expressed in GDP per capita to compare the wealth of different nations. http://www.gapminder.org/ for example uses it. The CIA Worldbook does so as well.

If the economics literature really does things different this seems to be a neat way that the CIA is actually using to push it's policy agenda.

If you have nearly data over the median income of countries all over the world over the last 100 years I would be interested in the data set.

I see a lot of data expressed in GDP per capita

GDP per capita is the aggregated Gross Domestic Product of the entire country divided by its population. It says nothing about income or wealth distribution within this country. It is also NOT the mean of personal income in that country.

GDP per capita basically tells you how much does a country produce, normalized for its population.

a neat way that the CIA is actually using to push it's policy agenda

If you have nearly data over the median income of countries all over the world over the last 100 years

I don't and I doubt it exists. You can find estimates of median income for developed countries during the last couple of decades easily enough in the usual places, but beyond that the data is likely to be sparse to absent.

I don't and I doubt it exists.

If you don't have the data how would you go about comparing the wealth of different countries based on it? I don't see how those claims fit together.

how would you go about comparing the wealth of different countries

I wouldn't. At least until the concept of "wealth of a country" gets defined.

But to answer your question, reread the last paragraph, particularly the part which starts "you can find...".

While the basic fact is that Christian is wrong, I think your response is an over-reaction. For example, the main point of PPP is to use GDP per capita as a typical income.

the main point of PPP is to use GDP per capita as a typical income.

Why do you think so? As far as I am concerned, the main point of PPP is to adjust the FX rates to make comparisons (of incomes, costs, living standards, etc.) between countries more meaningful.

PPP has nothing to do with the relationship between GDP per capita and personal (or household) income.

Yes, PPP is logically independent of GDP, but PPP GDP per capita is quite popular, probably the most common use of PPP.

Exactly. Someone succeeded in making the most common use of PPP something for which it's logically inappropriate. Of course it can be an accident of history but it's still interesting.

If your goal is to understand the difference between median and mean, developing sensitivity for issues like that is the point. Even if it's all accidents of history, you want to have issues like that get your conscious attention when looking at statistics.

The fact that the CIA factbook happens to be the best source of knowledge and the CIA is an organisation that heavily invest into shapping public discourse is something that can raise a bit of attention.

Of course it can be just because some bureaucrats inside the CIA are stupid that they don't give you the median income, because the have never thought that anyone would be interested in the median income. A much better explanation of why they don't give you that data is that they actually want that the GDP per capita is used that way. That people use GDP per capita when they want to think of the typical income of a country.

If people don't have any good data of the median income of countries, than that's what happens. Because of decisions by the CIA GDP per capita data is available but median income of countries isn't.

Statisticians in the CIA aren't stupid. They should at least be aware of the effects of choosing to report the wealth of countries that way.

The CIA also takes information warfare seriously. I live in Germany. After WWII there was a lot of investment in shaping German public opinion by the CIA. In the West people would be upset if the CIA world factbook gives them wrong numbers but nobody is upset if they are just given a number like GDP per capita when they want to know about the typical income.

Someone succeeded in making the most common use of PPP something for which it's logically inappropriate.

It is not the "most common use of PPP", at least in my neck of the woods.

the CIA factbook happens to be the best source of knowledge

It is not, not even close.

the CIA is an organisation that heavily invest into shapping public discourse

It is not.

Because of decisions by the CIA GDP per capita data is available but median income of countries isn't.

Really..? Are you posting drunk or something?

Sometimes that communist thought got in the way for cases where the arithmetic mean was really appropriate.

What is an example where the mean is better than the median?

I care more about my daily mean income in the last year than my median income over the last year.

Do you want to learn to do statistical analysis with a tool like R and interpret data? Do you want to learn mathematical axioms and theorems about probability and statistics and how to prove them.

Ideally, I'd like to learn both.

Depending on who you ask it's not. sum/n is the arithmetic mean. There are also other mean's like the geometric mean and the harmonic mean. Depending on the context different mean's can be used.

I know about different means. And I know that sometimes mean isn't appropriate at all (bimodal...) The context is one of the things I'd like to understand.

When it comes to learning notation. Anki is quite good.

Just used it to learn the Greek alphabet. :-)

Which stimulants/eugeroics have a short (< 3 hours) half-life? I did some research into this. Nicotine and selegiline (~1.5 hours) are the shortest I could find. Methylphenidate comes in next (~3.5 hours), but that's longer than I'd like. I don't particularly like any of these choices for various reasons and am interested in learning about others. Alternatively, if there's a way to significantly reduce the half-life of modafinil, I'd like to hear about that.

I've considered amphetamine, armodafinil, atomoxetine, caffeine, ephedrine, methylphenidate, modafinil, nicotine, pseudoephedrine, and selegiline.

Why do you care about the half-life?

Seems relevant for not interfering with sleep. For example, I can't use modafinil after 11AM because it interferes with my sleep that night if taken later; if I take modafinil or armodafinil at 5PM I might as well just skip the night's sleep. On the other hand, I can use caffeine up to 7/8PM without issue, and nicotine up to 10/11PM. (This is unfortunate because I am a bit of an owl and the evening is precisely when I'd like to be able to use a stimulant.)

Have you tried any drugs to fall asleep faster when using stimulants in the evening?

I don't usually look at interactions, because it's more important to establish an effect exists in the first place. Right now, I know that melatonin helps in getting to sleep, vitamin D impedes getting to sleep, high doses of magnesium citrate have ambiguous effects on sleep, alcohol usage seems to correlate with early bedtime (I forget about getting to sleep), and I speculate that Redshift/f.lux and masturbation help in getting to sleep but I haven't analyzed that experiment yet. I haven't looked at any interactions with nicotine or modafinil or anything, since I'd expect them to just independently make it harder to get to sleep.

If that's the goal why don't ask for it directly?

Ask for what directly? If he asked for stimulants that didn't interfere with sleep, he'd get replies suggesting... he look for ones with short half-lives like nicotine and avoid long-acting ones like modafinil. Short half-lives is the governing criterion, so he simply asked for it.

I think there are plenty of people that have tested various stimulants and know the effects those substances have on them, but who don't know the exact half-life of the substances.

I'm not even sure whether the real half life of caffeine is the thing that matters. When it comes to designing drugs there are quite a few things that are done in the delivery mechanism of the drug that can effect half-life.

Pharmacokinetics (half-life and other variables) of drugs and their different delivery methods are public knowledge, there are individual differences in metabolism of course.

When it comes to designing drugs there are quite a few things that are done in the delivery mechanism of the drug that can effect half-life.

For most drugs the elimination half-life is so long that a faster route of administration makes minor difference. For caffeine for example it's about 4.5 hours. You can shorten the absorption from < 1 hour to seconds but that probably won't matter much in this case.

You can make the absorption slower in a way that makes a difference however, for example there are several timed release versions of methylphenidate.

Pharmacokinetics (half-life and other variables) of drugs and their different delivery methods are public knowledge, there are individual differences in metabolism of course.

There are number that are public knowledge. Those numbers do have meaning. On the other hand individual differences are also important.

A lot of pseudoscience comes from people having a rough theory about isolated facts and personal observations. I think it's often very valuable to stay with personal observations instead of trying to fit them into a simple theory you made up that seems to fit and that corresponds to isolated facts you find in books.

You lose relevant information when you ask for half-life data instead of asking about what stimulants other people found useful for being stimulated at events closer to bed time without adverse effects on sleep.

Take Vitamin D. A lot of the published research on it is misguided because it presumes that blood level of Vitamin D is the central variable that matters. Whether you take Vitamin D in the morning or evening doesn't have much effect on long term Vitamin D blood levels. It therefore isn't subject to study in academia. Making too much assumption when you don't need to do so often hurts understanding.

Take Vitamin D. A lot of the published research on it is misguided because it presumes that blood level of Vitamin D is the central variable that matters. Whether you take Vitamin D in the morning or evening doesn't have much effect on long term Vitamin D blood levels. It therefore isn't subject to study in academia.

So, are you claiming that the blood level of vitamin D is NOT the central variable that matters? And that whether you take it in the mornings or in the evenings is important?

A bunch of QS people have observed that the timing of vitamin D supplements matters a great deal. Seth Robert wrote a lot about it (http://blog.sethroberts.net/2012/11/01/vitamin-d3-in-morning-improves-sleep-after-all-story-26/ for example).

As far as people you might trust, Gwern replicated the finding: http://www.gwern.net/Zeo#vitamin-d . Taking the supplement at night damaged his sleep.

I don't want to claim that blood level of vitamin doesn't matter at all, but I do claim that it's very unfortunate that there aren't more studies tracking the timing of vitamin D ingestion. I'm also thinking that getting the timing wrong is a good explantion for the studies that are out there that don't show improvement given vitamin D supplemention. Those studies are also the reason why the RDA of vitamin D is at 600 UI while QS folk generally recommend 2000 UI+ (again I think Gwern takes something like 5000 UI).

There a long term study called the VITAL study by Harvard Medical School in progress that tests the effects of 2000 UI vitamin D supplements on mortality rates. Unfortunately it doesn't track the intake of the timing so the resulting data might be worthless.

From the studies that showed effects for vitamin D you could deduce that the supplements can bring 2 additional years of lifespan. If the studies that say vitamin D does nothing come to that conclusion because of bad timing, that's a serious issue. As far as references go I remember the number of 3 years of lifespan for curing cancer.

I don't claim that I know with 100% certainity that it's in the timing and not in the average blood level. I do claim that the medical establishment is stupid for assuming that it's in the blood level. That's not even a real outsider opinion. That more or less the opinion I was thought by my bioinformatics professors. People in medicine make a lot of stupid assumptions that aren't based on evidence.

Now when I do QS I do make a bunch of assumptions that wouldn't pass in the academic context of bioinformatics. On of the great things about QS is that you aren't blind. You know reasonably well when you take your supplements while the doctors who administer clinical trials don't have any information in their data by default about when their subjects take the supplements.

As a result it's good practice to stay near empiricial data and not make assumption unless they will help you.

Model the problem as seeking of being tired as seeking an efficient stimuluant is a choice. When I instead propose that he should focus on getting better at relaxing than I'm pushing a different model for the situation.

It's not that my model is inherently based on the truth. It is in some sense "science inspired" when I use the mental model of the body downregulating itself via a cybernetic process. Thinking in terms of cybernetics (the word doesn't get used much these days) is one of the model I learned at university. That doesn't make it right but it's an available model to explore for the problem.

Now I have different kind of evidence that over a handful of different trance states that I learned about in different contexts help people to sleep better afterwards. Hypnotherapists do have a body of theory that predicts that's a usual effect side effect of hypnosis. One of my hypnosis teachers for example told a story about how a person who didn't even spoke his language and who was escorting a patient got into trance while watching the session and afterwards resolved her problem of not being able to sleep well. It's an effect that can happen as "correlateral damage".

I don't know the exact kind of relaxation protocol that best for btrettel. I didn't even try to push my favorite relaxation protocol that gave me QS validated benefits for another medical issue because he wouldn't find a practioner for it in Texas anyway.

Comparing different relaxation protocols against each other isn't something that well done by the academic establishment because it not really in the model of how to go about treating a patient. No patents that pay for expansive clinical trials. Exactly the area where it's good to do your your empirics.

A bunch of QS people have observed that the timing of vitamin D supplements matters a great deal.

Matters for what?

Your links say that the timing of vitamin D intake affects sleep. Fine, but that's not really what most people take vitamin D for. There is a variety of claims for vitamin D supplements which generally have to do with bone health, viral infections, CVD, etc. I don't want to get into evidence for and against these claims, but are you saying that the timing of vitamin D affects these outcomes?

Looking at it in the most general fashion, the overall claim is that taking vitamin D supplements affect mortality. Crudely, you live longer. That may or may not be so, but do you think that timing of vitamin D ingestion would affect that? What evidence do you have? Sleep disturbance is a time-local short-term effect, it isn't obvious to me that it indicates problems with long-term consequences.

I do claim that the medical establishment is stupid for assuming that it's in the blood level.

Why? "Stupid" is a strong word. If your hypothesis is that timing matters but the blood level doesn't matter, what's the underlying biochemical mechanism? Is there any evidence that the right time in the circadian cycle is crucial?

Also, if you are taking vitamin D supplements wouldn't you be interested in your blood level? How would you know how much of vitamin D do you need?

If your hypothesis is that timing matters but the blood level doesn't matter, what's the underlying biochemical mechanism? Is there any evidence that the right time in the circadian cycle is crucial?

I see evidence that both matter to sleep. It seems that the blood level of vitamin D is linked to excessive daytime sleepiness. (Warning: the authors of this paper overfit the data, so you can ignore their conclusions about race and very low vit. D levels, but their data does show a negative correlation between serum vitamin D levels and excessive daytime sleepiness.)

My own serum vitamin D level was pretty low, but since supplementation, it has increased appreciably to within the normal range. So far, I don't think it has had an effect on my daytime sleepiness, but I have not been keeping track of the appropriate factors, so take what you will.

It also seems that taking vitamin D at night seems to disrupt sleep for some individuals. My experience suggests taking vitamin D at night has no effect, but (as before) I have no hard data to justify this. It is possible that the sleep disruption only applies to those who have adequate blood levels of vitamin D. The explanation that I have seen (which I can't find right now) is that vitamin D influences your circadian drive as sunlight would because your body synthesizes it from sunlight; taking vitamin D is like getting "concentrated sunlight". I'll agree with ChristianKI, though, that no mechanism needs to be identified to validate an observation.

For other things (i.e., not sleep), I haven't seen any evidence of timing effects.

It seems that the blood level of vitamin D is linked to excessive daytime sleepiness. (Warning: the authors of this paper overfit the data, so you can ignore their conclusions about race and very low vit. D levels, but their data does show a negative correlation between serum vitamin D levels and excessive daytime sleepiness.)

I haven't read the paper, just looked at their plots, and my impression is that there is nothing there but noise.

It also seems that taking vitamin D at night seems to disrupt sleep for some individuals.

We have anecdata, but have there been actual studies?

And speaking of timing of vitamin D supplementation, it is well-known that the absorption of it varies, in particularly depending on whether you take it with fats (in your food) or not. That would have to be controlled for in any experiments designed to figure out timing effects.

I haven't read the paper, just looked at their plots, and my impression is that there is nothing there but noise.

That may be true. The correlation is at best weak. There appears to not necessarily be a causative link between vit. D and daytime sleepiness; increasing my vit. D levels had no perceptible effect on my own sleepiness. Though others have had different experiences.

We have anecdata, but have there been actual studies?

I have not seen any studies into that. The closest that I've seen is gwern's tests.

Sleep disturbance is a time-local short-term effect, it isn't obvious to me that it indicates problems with long-term consequences.

Why do you think humans sleep at all if sleep disturbance has no long term effects? I think it's fairly straightforward to think that humans do undergo processes that further health during restful sleep. After quick Googling http://www.ncbi.nlm.nih.gov/books/NBK19961/ is a study that says so.

Your links say that the timing of vitamin D intake affects sleep. Fine, but that's not really what most people take vitamin D for. There is a variety of claims for vitamin D supplements which generally have to do with bone health, viral infections, CVD, etc.

Bone health might be just about Vitamin D's role in calcium absorption.

From the paper I linked above:

Sleep Loss Is Associated with Cardiovascular Morbidity

Sleep loss and sleep complaints are associated with heart attacks (myocardial infarction) and perhaps stroke, according to several large epidemiological studies

I can't find talk about viral infections on that page but I would assume that you can also make a case that a sleep deprived individual is at higher risk for them.

Why? "Stupid" is a strong word.

Investing tens of millions in experiments based on a hypothesis that you don't really test is stupid. To use the words of Feynman you could also say cargo cult science with Feynman used to describe the rat psychology experiments of his time.

If your hypothesis is that timing matters but the blood level doesn't matter, what's the underlying biochemical mechanism?

When I say blood level I mean the level you measure when you give a individual a blood test every month and make a study based on that data. I don't mean the level you would get if you measure every minute.

But I don't need to point to a biochemical mechanism to validate an empirical observation. Currently drugs get often designed based on an idea that you want to target a biochemical mechanism but when they do work, the work in mysterious ways that aren't exactly the way the people who designed the drug would have thought beforehand. Of course most of those drugs fail anyway. It's much better to focus on things that produce empiric effects than going to deeply into theory.

But as far as vitamin D goes, there plenty of evidence that it can work as a hormone. It also a hormone that gets naturally produced at specific times as the sun usually shines at specific times of the day and not at night.

How would you know how much of vitamin D do you need?

The empirical method. You can take different amount of vitamin D and see the effect on yourself. That means you have either good awareness of your own body, QS tools or both.

Coming to your own judgments instead of trying to follow what some authoritative doctor or doctrine tells you is what Kant described in his day's as his ideal of enlightenment. The way is real empiricism. Paying attention to real world feedback.

Investing tens of millions in experiments based on a hypothesis that you don't really test is stupid.

The hypothesis being tested is that the blood level of vitamin D is relevant for the outcomes. You think they should test another hypothesis but that doesn't mean the original researchers are stupid.

The empirical method. You can take different amount of vitamin D and see the effect on yourself.

If I am interested in the effect of vitamin D on overall mortality, it's kinda difficult to "see the effect on [my]self".

Coming to your own judgments instead of trying to follow what some authoritative doctor or doctrine tells you is what Kant described in his day's as his ideal of enlightenment.

Yes, but you're confused between blindly following authority and looking at data from people other than yourself.

The hypothesis being tested is that the blood level of vitamin D is relevant for the outcomes. You think they should test another hypothesis but that doesn't mean the original researchers are stupid.

I didn't call individuals stupid but I spoke about the practice they follow. I also don't call 18st century scientists stupid even when a lot of their mental models were stupid from the perspective of knowing what I know today.

In this case, before you spend a lot of money on a long term mortality study it's better to run a few smaller studies to gauge whether variables such as the timing have an effect.

If I am interested in the effect of vitamin D on overall mortality, it's kinda difficult to "see the effect on [my]self".

Until the VITAL study get's completed it's also impossible to get that data elsewhere directly. Just that you don't misunderstand myself, I don't oppose that fact that the VITAL study get's run. It's better value for money than many other things nutrition academics fund.

I mean at the moment we have the situation that we do have a meta review that says that we can expect to gain two years of life expectancy via daily 2000 UI vitamin D supplements.

We have other academics that are less optimistic. But nearly nobody claims that taking 2000 UI vitamin D is really dangerous. Academics have different opinions on whether you should take vitamin D supplements.

Additionally you don't lose anything as an individual if you take your vitamin D in the morning because of anecdotal evidence. Even if the timing doesn't matter you still get the benefit.

Yes, but you're confused between blindly following authority and looking at data from people other than yourself.

I never said that one shouldn't look at data from people other than yourself. I said you shouldn't simply copy their way of modeling the problem. Even when it comes to something like hypnosis/NLP I'm perfectly willing to read academic papers and try to understand the empirical observations that they made. I might not agree with the interpretation but I'm not one to turn down good data.

There no good data at all for the claim that taking blood vitamin measurements and changing the amount of vitamin D supplements that you consume based on that data does anything for you that's better than just taking 2000UI (or 5000UI). That not something that they studied as far as I knowledge is concerned.

Kant was explicit in his papers that one shouldn't use his doctor has authority for one's health to override your own self determination.

Search on vitamin D at blog.sethroberts.net -- you'll get a bunch from the self-tracking perspective.

gwern is partly right about my motivations.

I have some sort of sleep disorder (perhaps narcolepsy; the sleep study I had was inconclusive on that, though it did rule out sleep apnea) where the treatment might involve taking stimulants. My sleep quality at night might be particularly bad in general, and taking a stimulant could make it worse. Modafinil's half-life is about 15 hours, which is quite long. Even if I took it in the morning, it still might impact my sleep at night.

Also, the stimulants I have had generally do not agree well with me. Often they make me nervous, even in low doses.

If I were to take a stimulant regularly, I'd rather take a short-acting one, and only take it when absolutely necessary.

I have some armodafinil now and am going to do a study on myself. My sleep doctor suggests testing my concerns as they may not pan out in reality.