Monthly Archives: August 2012

Surplus splitting strategy

When negotiating over the price of a nice chair at a garage sale, it can be useful to demonstrate there is only twenty dollars in your wallet. When determining whether your friend will make you a separate meal or you will eat something less preferable, it can be useful to have a longterm commitment to vegetarianism. In all sorts of situations where a valuable trade is to be made, but the distribution of the net benefits between the traders is yet to be determined, it can be good to have your hands tied.

If you can’t have your hands tied, the next best thing is to have a salient place to split the benefits. The garage sale owner did this when he put a price tag on the chair. If you want to pay something other than the price on the tag, you have to come up with some kind of reason, such as a credible commitment to not paying over $20. Many buyers will just pay the asking price.

This means manipulating salient ways to split benefits could be pretty profitable. This means people should probably be doing it on purpose. I’m curious to know if and how they do.

Often the default is to keep the way the benefits naturally fall without money (or anything else ‘extra’) changing hands. For instance suppose you come to lunch at my place and we both enjoy this to some extent. The default here is to keep the happiness we got from this, rather than say me paying you $10 on top.

So in such cases manipulating the division of benefits should mostly be done by steering toward more personally favorable variations on the basic plan. e.g. my suggesting you come to my place before you suggest that I come to yours. A straightforward way to get gains here is to just race to be the first to suggest a favorable option, but this is hard because it looks domineering to try to manipulate things in your favor in such a way. Unless you have some particular advantage at suggesting things fast and smoothly, such a race seems costly in expectation.

If in general trying to manipulate a group’s choice seems like a status-move or dominance-move, subtle ways to do this are valuable. Instead of a race to suggest options, you can have a prior race to make the options that you might want to suggest seem more suggestible. For instance if you’d prefer others come to your place than you go to others’ places, you can put a pool at your place, so suggestions to go to your place seem like altruism. If you know a lot of details about another person, you can use one of them to justify assuming that a particular outcome will be better for them. e.g. ‘We all know how much John likes steak, so we could hardly not go to Sozzy’s steak sauna!’. None of this works unless it’s ambiguous which way your own preferences go.

On the other hand if your preferences are very unambiguous, you can also do well. This is because others know your preferences without your having to execute a dominance move to inform them. If their preferences are less clear, it’s hard for them to compete with yours without contesting your status themselves. So arranging for others to know your preferences some other way could be strategic. e.g. If you and I are choosing which dessert to split, and it is common knowledge that I consider chocolate cake to be the high point of human experience, it is unlikely that we will get the carrot cake, even if you prefer it quite strongly.

So, strategy: if it’s clear that you have a pretty strong preference, make it quite obvious but not explicit. If you have a less clear preference, make it look like you have no preference, then position to get the thing you want based on apparently irrelevant considerations.

Even if the default is to transfer no cash, there can be a range of options that are clearly incrementally better for you and worse for me, with no salient division. e.g. If I invite you over for lunch, there are a range of foods I could offer you, some better for you, some cheaper for me. This seems quite similar to determining how much money to pay, given that someone will pay something.

In the lunch case I get to decide how good what I offer you is, and you have to take it or leave it. You can retaliate by thinking better or worse of me. You can’t very explicitly tell me how much you will think better or worse of me though, and you probably have little control over it. Your interpretation of my level of generosity toward you (and thus your feelings) and my expectations of your feelings are both heavily influenced by relevant social norms. So it’s not clear that either of us has much influence over which point is chosen. You could try to seem unforgiving or I could try to seem unusually ascetic, but these have many other effects, so are extreme ways to procure better lunching deals. I suspect this equilibrium is unusually hard to influence personally because there’s basically no explicit communication.

There are then cases where money or peanut butter sandwiches or something does change hands naturally, so ‘no transfer’ is not a natural option. Sometimes there is another default, such as the cost of procuring whatever is being traded. By default businesses put prices on items rather than consumers doing it, which appears to be an issue of convenience. If it’s clear how much surplus is being split, a natural way is to split it evenly. For instance if you and I make $20 busking in the street, it would be strange for you to take more than $10, even if you are a better singer. This fairness norm is again hard to manipulate personally, except by making it more or less salient. But it’s a nice example of a large scale human project to alter default surplus division.

When there are different norms among different groups, you can potentially reap more of it by changing groups. e.g. if you are a poor woman, you might do better in circles where men are expected to pay for many things.

These are just a random bunch of considerations that spring to mind. Do you notice people trying to manipulate default surplus divisions? How?

GD Star Rating
Tagged as: , , ,

Sex At Dusk v. Dawn

Two years ago I was persuaded by the book Sex At Dawn, at least on its “key claim, that forager females were sexually promiscuous.” While I didn’t buy authors’ free-love scenario, I thought our ancestors were much less tied to their sex partners than most folks realize:

A Hadza man hunts big game to look sexy, even though that retrieves less food. Except that when a women he has sex with has a kid he thinks is his, he’ll gather more but less-sexy food, to give this woman ~1/2 of her food for one year, ~1/4 for the next two years, and declining amounts thereafter. Now, yes, this may be more pair-bonding than in chimps or bonobos. But it is also far less than the farmer ideal of life-long monogamy! Many men today reluctant to marry for life would be ok with this level of commitment.

Lynn Saxon has a new book Sex At Dusk, quite critical of Sex At Dawn. She was a kind enough sent me a copy, which I’ve just read. Searching, I’ve only find positive reviews of it (here, here, herehere).

Saxon doesn’t write as well, and especially fails to summarize well. Even so, she does successfully undercut many Sex At Dawn arguments. In humans, sexual jealousy is a universal, females are picky about sex partners, penises aren’t over-sized, testes are small, sperm production slow, and the evidence doesn’t suggest a great deal of sperm competition. Female chimps have little extra-group sex, bonobos don’t usually mate face-to-face, and many Sex At Dawn quotes are misleading, given their context.

On sound during sex, Saxon offers evidence that female primate cries during sex aren’t simple invitations:

A recent study of female chimpanzees found no support for the sperm competition hypothesis: females did not produce calls when mating with low-ranked males so it was not about inviting other males to join the party, and calls did not correlate with fertility and the likelihood of conception. … Females called significantly more while mating with high-ranked males, but suppressed their calls if high-ranked females were nearby. …

[Researchers] found that [bonobo] females were more likely to call with male rather than female partners but the patterns of call usage were very similar in that females called more with high ranked partners (as in chimpanzees), regardless of the partner’s sex. With a female partner compilation calls were consistently produces only by the lower ranking of the two females. … In bonobos the increase in calls during the alpha female’s presense. (p.279-280)

Yes, it looks like chimp and bonobo sex calls are more brags than invitations. Even so, brags make little sense when it is common knowledge who has sex with whom. So a habit of similar bragging sex calls by human females would suggest that humans often didn’t know who was having sex with whom, suggesting a lot of promiscuity.

A key question, to me, is what percentage of our forager ancestor kids were fathered outside pair-bonds. That is, what fraction of kids were born to mothers without a main male partner, or had a father different from that partner. This number says a lot about the adaptive pressures our ancestors experienced related to various promiscuous and polyamorous arrangements today. And hence says a lot about how “natural” are such things.

Alas, none of these authors give a number, but my impression was that Saxon would estimate less than 20%, while the Sex At Dawn authors would estimate over 50%. Even 20% would be consistent with a lot of human promiscuity adaptations, such as female sex brag calls. I asked Saxon directly via email, however, and she declined to give a number – she says her main focus was to argue against Sex At Dawns‘ “paternity indifference” theory (that humans don’t care which kids are theirs). Which is fine – that is indeed a pretty crazy theory.

So where does the evidence sit on promiscuity? Our closest living relatives, chimps and bonobos, are quite promiscuous. Yes, their pair bonds much weaker than ours, and pair-bonding usually greatly reduces promiscuity. But few pair-bonded animals live in big social groups where hidden extra-pair sex is so easy to arrange, and humans live in even bigger groups than chimps or bonobos. For humans, we have lots of clear evidence of outside-pair sex, mate-guarding to prevent such, bragging sex cries, and desires for sexual variety. And humans do seem to spend a record fraction of their time thinking about and doing sex.

Since humans usually have clear overt norms against extra-pair sex, but strong urges to arrange covert sex, promiscuity estimates comes down in part to estimates of how well humans actually enforced their norms. My homo hypocritus theory suggests a lot of covert norm violation, and so a lot of promiscuity. So I’ll estimate the key promiscuity number in the 20-30% range.

Btw, here’s another fascinating quote from the book:

Hrdy writes about a startling interview with an old hunter in which he reminisces to a time when just the sound of his footsteps on the leaves of the forest floor struck terror in the hearts of old women. He was the socially sanctioned specialist in eliminating old women deemed no longer useful: coming up behind an old woman he would strike her on the head with his axe. (p.216)

GD Star Rating
Tagged as: , ,

How To Vote

I’m a professor of economics who has published on politics, and in ten weeks the US will elect a president, and congress-folk. So a reasonable test of my supposed specialized knowledge is: do I have anything clearly useful to tell ordinary folks about how to vote in this election?

Well first, I have said that politics isn’t about policy, i.e., that we quite reasonably care more about how our political views make us look to our associates than how they effect policy outcomes. I’ve also said a lot about overconfidence, which applies in spades to politics. Being overconfident usually looks better, even it leads to less accurate judgements. So in general I can recommend: just confidently take positions that look good to your associates; forget about policy outcomes.

But what if you were to insist on trying to figure out which US presidential and Congressional candidates to vote for based on expected policy outcomes? And what if you were to further insist on avoiding overconfidence, seeking as much as possible to avoid relying on your own likely-overconfident judgements. Perhaps, for example, the image you want to project to your associates is of fair-minded broad concern, and well-calibrated rationality.

In this case, if you guess that your relevant information is roughly average or worse than average, you have an obvious solution: don’t vote. But, perhaps you want to signal a bit more confidence and concern than this suggests.

Another simple strategy is to search for an immediate associate who both clearly shares most of your interests, and who seems better informed than you. Then just copy their vote. But perhaps you don’t want to appear very submissive to this person. So let us assume that you want to signal that you are informed and autonomous, while still creating good voting outcomes, and continue.

Alas, we still don’t have presidential decision markets, that estimate important policy outcomes, such as GDP, unemployment, oil prices, etc., conditional on who becomes president. It would only take a few tens of thousands of dollars for someone to create these, at say Intrade. And then I could offer great simple clear robust advice – vote for whomever markets rate better. But, ok, this system failing is not your problem, so let’s continue.

The thing you probably know best is your own life. So a good simple strategy is to vote “retrospectively,” i.e., for incumbents if your life goes well, and against them if your life goes badly. The more voters who do this, the stronger incentives incumbents have to make most lives go well.

For life quality extremes this advice is clear, but what if your life is near the middle? What should be the cutoff between a good and bad life? One simple reference is how you expected your life to go when incumbents were elected – reelect them if your life has gone better than expected.

Now you might do a little better if you could broaden your judgment to how life has gone for more people you care about. But if you care about a lot of distant folks, you’ll face the problem that you know a lot less about how distant folk lives are going. I could just tell you that most experts agree that the US economy has done worse than experts expected when Obama was elected. But to rely on that, you might have to trust me.

You could also do a bit better if you focused on the outcomes where particular incumbents are most responsible. For example, US presidents have more influence on foreign policy, while Congress has more influence on domestic policy. US politicians have more effect on what happens in the US than in Europe. Politicians have less influence on earthquakes or hurricanes. And so on. You might also substitute the initial expectation you should have had, for the one you did have, if you can figure out that. But you have to know more than most people do to implement such corrections well, and they only moderately improve incumbent incentives.

So, as a professor of economics who has studied politics, my advice is to not vote if you know an average amount or less, to copy a better informed close associate if you are willing to appear submissive, and otherwise to just reelect incumbents when your life goes better than you expected. And if you care a lot more about the outcome than most do, help create presidential decision markets, so other info-seekers will have a better place to turn.

GD Star Rating
Tagged as: , ,

AI Progress Estimate

From ’85 to ’93 I was an AI researcher, first at Lockheed AI Center, then at the NASA Ames AI group. In ’91 I presented at IJCAI, the main international AI conference, on a probability related paper. Back then this was radical – one questioner at my talk asked “How can this be AI, since it uses math?” Probability specialists created their own AI conference UAI, to have a place to publish.

Today probability math is well accepted in AI. The long AI battle between the neats and scruffs was won handily by the neats – math and theory are very accepted today. UAI is still around though, and a week ago I presented another probability related paper there (slides, audio), on our combo prediction market algorithm. And listening to all the others talks at the conference let me reflect on the state of the field, and its progress in the last 21 years.

Overall I can’t complain much about emphasis. I saw roughly the right mix of theory vs. application, of general vs. specific results, etc. I doubt the field would progress more than a factor of two faster if such parameters were exactly optimized. The most impressive demo I saw was Video In Sentences Out, an end-to-end integrated system for writing text summaries of simple videos. Their final test stats:

Human judges rated each video-sentence pair to assess whether the sentence was true of the video and whether it described a salient event depicted in that video. 26.7% (601/2247) of the video-sentence pairs were deemed to be true and 7.9% (178/2247) of the video-sentence pairs were deemed to be salient.

This is actually pretty impressive, once you understand just how hard the problem is. Yes, we have a long way to go, but are making steady progress.

So how far have we come in last twenty years, compared to how far we have to go to reach human level abilities? I’d guess that relative to the starting point of our abilities of twenty years ago, we’ve come about 5-10% of the distance toward human level abilities. At least in probability-related areas, which I’ve known best. I’d also say there hasn’t been noticeable acceleration over that time. Over a thirty year period, it is even fair to say there has been deceleration, since Pearl’s classic ’88 book was such a big advance.Robin Hanson

I asked a few other folks at UAI who had been in the field for twenty years to estimate the same things, and they roughly agreed – about 5-10% of the distance has been covered in that time, without noticeable acceleration. It would be useful to survey senior experts in other areas of AI, to get related estimates for their areas. If this 5-10% estimate is typical, as I suspect it is, then an outside view calculation suggests we probably have at least a century to go, and maybe a great many centuries, at current rates of progress.

Added 21Oct: At the recent Singularity Summit, I asked speaker Melanie Mitchell to estimate how far we’ve come in her field of analogical reasoning in the last twenty years. She estimated 5 percent of the way to human level abilities, with no noticeable acceleration.

Added 11Dec: At the Artificial General Intelligence conference, Murray Shanahan says that looking at his twenty years experience in the knowledge representation field, he estimates we have come 10% of the way, with no noticeable acceleration.

Added 4Oct’13: At an NSF workshop on social computing, Wendy Hall said that in her twenty years in computer-assisted training, we’ve moved less than 1% of the way to human level abilities. Claire Cardie said that in her twenty years in natural language processing, we’ve come 20% of the way. Boi Faltings says that in his field of solving constraint satisfaction problems, they were past human level abilities twenty years ago, and are even further past that today.

Let me clarify that I mean to ask people about progress in a field of AI as it was conceived twenty years ago. Looking backward one can define areas in which we’ve made great progress. But to avoid selection biases, I want my survey to focus on areas as they were defined back then.

Added 21May’14: At a private event, after Aaron Dollar talked on robotics, he told me that in twenty years we’ve come less than 1% of the distance to human level abilities in his subfield of robotic grasping manipulation. But he has seen noticeable acceleration over that time.

Added 28Aug’14: After coming to a talk of mine, Peter Norvig told me that he agrees with both Claire Cardie and Boi Faltings, that on speech recognition and machine translation we’ve gone from not usable to usable in 20 years, though we still have far to go on deeper question answering, and for retrieving a fact or page that is relevant to a search query we’ve far surpassed human ability in recall and do pretty well on precision.

Added 14Sep’14: At a closed academic workshop, Timothy Meese, who researches early vision processing in humans, told me he estimates about 5% progress in his field in the last 20 years, with a noticeable deceleration.

Added 4Jan’15: At a closed meeting, Francesca Rossi, expert in constraint reasoning, gave an estimate of 10%, with deceleration. Margret Boden, author of Artificial Intelligence and Natural Man (1977), estimated 5%, but for no particular subfield.

Added 6July’15: David Kelley, expert in big data analysis, says 5% in last twenty years, sees acceleration only in last 2-3 years, not before that.

Added 18Apr’16: Henry Kautz, says in constraint satisfaction we were at human level 20 years ago and have moved to super human levels now. In language, he says we’ve moved 10% of the way, with a noticeable acceleration in the last five years.

Added 13July2016: Jeff Legault says that in robotics we’ve come 5% of the way in the last 20 years, and there was only acceleration in the last five years.

Added 08Sept2017: Thore Husfeldt says that in the field of human understandable explanation, we have come less than 0.5% of the distance.

GD Star Rating
Tagged as: , ,

Fairy Tales Were Cynical

A recent New Yorker article on fairy tales fascinated me (quotes below). Apparently the fairy tales once “told at rural firesides” were for adults, full of sex and violence, and cynical – they did not often affirm common ideals. This stands in sharp contrast to most fiction genres today, especially today’s fairy tales targeted at kids. Why were long ago stories so much more cynical? They remind me of some joke genres, like dead baby jokes, and of the crudeness often found off the record in many close social groups.

Here’s my homo hypocritus explanation. Our forager ancestors evolved intricate capacities to affirm standard ideals when what they said or did might be visible or reported to distant observers, and to coordinate to violate such ideals when they were less visible. Shared private rejection and violation of wider ideals can signal close bonds with associates, and reveal more about ourselves to intimates.

So when stories become more visible, such as by getting published in books, stories had to become more ideal. Similarly, when kids were taught in schools, with a curriculum visible to all, that curriculum had to become more ideal. And as law enforcement has become more visible, it has been held to higher standards.

Today harassment laws make it harder to be very crude and cynical at work, and divorce custody battles punish parents who act this way around their kids. Today, more interactions are governed by officially idealistic norms: teachers around students, doctors & lawyers around clients, etc. What costs do we pay for this panopticon-like suppression of our natural crude/cynical styles? We are probably less able to form very close social groups where we can more clearly see each others’ weaknesses and vulnerabilities. But what else?

Added 26Aug: Another contributing factor may be that in general our idealism just rises with rising wealth.

Those promised quotes: Continue reading "Fairy Tales Were Cynical" »

GD Star Rating
Tagged as: , ,

Middle-aged not miserable, just too busy to answer surveys

I have read several times that there is evidence of a U-curve in happiness over an individual’s life. People are happy in their youth, and happy again after retirement, but suffer from a serious malaise in between as they grapple with their finances, careers and family life. Seeing as how I’m about to embark on this part of my life, that isn’t a particularly appealing idea! Today I was glad to find some evidence that the U-curve is just a statistical illusion. Australian economists Paul Frijters and Tony Beatton have analysed large panel data sets from Australia (n>10,000), the UK (n>25,000) and Germany (n>20,000) and produced the following trajectory for happiness as people age:

Happiness is nearly flat from 20 through 50. Frijter’s explanation for the disagreement with the existing literature is that,

previous studies severely underestimated the degree to which miserable people in middle-age were over-represented in these datasets: happy people in middle age are busy and don’t have time to participate in lengthy questionnaires, leading previous studies to erroneously think there was a huge degree of unhappiness in middle-age. When you actually follow people over time, no such ‘middle-age blues pattern’ can be found, at least not in Australia or the UK and only to a mild degree in Germany. … we found that there were severe data problems in Germany, with only quite miserable people in middle-age prepared to partake in the sample and respondents becoming markedly more honest (and miserable) as they answered the happiness questions year after year.

Those of us with active lives need not fear! There are familiar selection problems later in life as well. Unhappy people do not live as long, and may become less able or willing to answer surveys than their healthy and cheerful counterparts. By tracking the same participants for 10 years or more, Frijters claims to have “dealt with both issues,” presumably suing their initial responses to by forecast their missing responses later in life. Some day I should choose carefully where I decide to retire and perhaps return to my homeland: “life in old age is clearly relatively better in Australia than the UK, perchance because of the better weather, more generous public pensions, and more space.”

Another, more depressing, paper on happiness recently caught my attention. Probably due to its early use of twins to investigate the causes of happiness it has landed over 900 citations, despite a limited sample of separated twins. It finds a remarkably high correlation between the happiness of twins who share all of their genes, but very little correlation for twins who share half of their genes:

Happiness or subjective wellbeing was measured on a birth-record based sample of several thousand middle-aged twins using the Well Being (WB) scale of the Multidimensional Personality Questionnaire (MPQ). Neither socioeconomic status (SES), educational attainment, family income, marital status, nor an indicant of religious commitment could account for more than about 3% of the variance in WB. From 44% to 53% of the variance in WB, however, is associated with genetic variation. Based on the retest of smaller samples of twins after intervals of 4.5 and 10 years, we estimate that the heritability of the stable component of subjective wellbeing approaches 80%.

For the 48 DZ [fraternal] pairs, this cross-twin, cross-time correlation for WB was essentially zero (.07) while, for the 79 MZ [identical] pairs, it equaled .40, or 80% of the retest correlation of .50. The MZ data suggest that the stable component of wellbeing (i.e., trait-happiness) is largely determined genetically. The negligible DZ correlation suggests that this stable and heritable component of happiness is an emergenic trait (Lykken, 1982; Lykken, Bouchard, McGue, & Tellegen, 1992), that is, a trait that is determined by a configural rather than an additive function of components. Emergenic traits, although determined in part genetically, do not tend to run in families as do traits that are polygenic-additive.

A similar result was reported in an earlier study of 217 MZ and 114 DZ pairs of middle-aged Minnesota Registry twins, plus 44 MZ and 27 DZ pairs who were separated in infancy and reared apart (Tellegen, et al., 1988). The best estimate of the heritability of WB in that study was .48 (± .08) and, as was true here, a model involving only additive genetic effects did not fit the data.

Myers and Diener suggested that people who enjoy close personal relationships, who become absorbed in their work, and who set themselves achievable goals and move toward them with determination are happier on the whole than people who do not. We agree, but we question the direction of the causal arrow. We know that when people with bipolar mood disored are depressed, they tend to avoid intimate encounters or new experiences and tend to brood upon depressing thoughts rather than concetrating on their work. Then, when their moods swings toward elation, these same people tend to do the things that happy people do. This is undoubtedly a James-Lange feedback effect: Dysfunctional behavior exacerbates depression, whereas the things happy people do enhance their happiness. We argue, however, that the impetus is greater from mood to behavior than in the reverse direction. It may be that trying to be happier is as futile as trying to be taller and therefore is counterproductive. (HT Jim Savage)

My impression – which I would be happy to have corrected – is that later research finds a smaller, but still significant impact of genetics on happiness.

That result brought to mind something John Stuart Mill wrote in his autobiography: “I am now convinced that no great improvements in the lot of mankind are possible until a great change takes place in the fundamental constitution of their modes of thought.” I imagine Mill had education and cultural shifts in mind, and the consistent difference in reported subjective wellbeing between South America and Eastern Europe show culture makes a difference. But if genetics plays such a large and limiting role, the only way to drastically alter our ‘modes of thought’ will require that we learn how to tinker with our minds directly.

GD Star Rating
Tagged as: , ,

More 2D Values

Back in ’09 I posted on the 2D map of values from the World Values Survey, and how nations are distributed in that 2D space. A related 2D space of values is detailed in this new JPSP paper. Apparently 19 different values fall naturally on a circle:

Here are more detailed descriptions of these values:

  • Self-direction–thought: Freedom to cultivate one’s own ideas and abilities
  • Self-direction–action: Freedom to determine one’s own actions
  • Stimulation: Excitement, novelty, and change
  • Hedonism: Pleasure and sensuous gratification
  • Achievement: Success according to social standards
  • Power–dominance: Power through exercising control over people
  • Power–resources: Power through control of material and social resources
  • Face: Security and power through maintaining one’s public image and avoiding
  • humiliation
  • Security–personal: Safety in one’s immediate environment
  • Security–societal: Safety and stability in the wider society
  • Tradition: Maintaining and preserving cultural, family, or religious traditions
  • Conformity–rules: Compliance with rules, laws, and formal obligations
  • Conformity–interpersonal: Avoidance of upsetting or harming other people
  • Humility: Recognizing one’s insignificance in the larger scheme of things
  • Benevolence–dependability: Being a reliable and trustworthy member of the ingroup
  • Benevolence–caring: Devotion to the welfare of ingroup members
  • Universalism–concern: Commitment to equality, justice, and protection for all people
  • Universalism–nature: Preservation of the natural environment
  • Universalism–tolerance: Acceptance and understanding of those who are different from oneself

Of course since they are based on surveys, these are probably mostly about values as seen in a far-view.

Added 21Aug: The upper values on the circle are those celebrated more by richer societies like ours, relative to poorer societies like our farmer ancestors. (Foragers were more in the middle.) In older societies, the upper values are also more celebrated by the rich. The left-side more-community-oriented are also more common in the “East,” which I’ve suggested were centrally located places more often conquered by invaders. The more peripheral “West” tended more to emphasize right-side family and individual values.

Added 24 Aug: Far mode emphasizes the positive over the negative, and the social over the personal. So the upper left area of the circle are the most far values, and the lower right the most near values. This also seems to map onto the (near) things that we actually want, and the (far) things we want others to think that we want.


GD Star Rating
Tagged as: ,

On the goodness of Beeminder improves my life a lot. This is surprising: few things improve my life much, and when they do it’s usually because I’m imagining it. Or because they are things that everyone has known about for ages and I am slow on the uptake (e.g. not moving house three times a year, making a habit of eating breakfast, making habits at all). But Beeminder is new, and it definitely helps.

One measurable instrumental benefit of Beeminder is that I have exercised for half an hour or an hour per day on average since last October. Previously I exercised if I needed to get somewhere or if the fact that exercise is good for people crossed my mind particularly forcibly, or if some even less common events occurred. So this is big. It seems to help a lot for other things too, such as working, but the evidence there is weaker since I used to work pretty often anyway. I’m sorry that  I didn’t keep better track.

Unlike many other improvements to my life, I have some guesses about why this is so useful. But first let me tell you the basic concept of Beeminder.

Take a thing you can measure, such as how many pages you have written. Suppose you measure this every day, and enter the data as points in a graph. Suppose also that the graph contains a ‘road’ stretching up ahead of your data, to days that have not yet happened. Then you could play a game of keeping your new data points above the road. A single day below the road and you lose. It turns out this can be a pretty compelling game. This is basically Beeminder.

There are more details. You can change the steepness of the road, but only for a week in the future. So you can fine-tune the challengingness of a goal, but can’t change it out of laziness unless you are particularly forward thinking about your laziness (in which case you probably won’t sign up for this).

There is a lot of leeway in what indicators you measure, and some I tried didn’t help much. The main things I measure lately are:

  • number of 20 minute blocks of time spent working. They have to be continuous, though a tiny bit of interruption is allowed if someone else causes it
  • time spent exercising weighted by the type of exercise e.g. running = 2x dancing = 2 x walking
  • points accrued for doing tasks on my to-do list. When I think of anything I want to do I put it on the list, whether it’s watching a certain movie or figuring out how to make the to do list system better. Some things stay there permanently, e.g. laundry. I assign each task a number of points, which goes up every Sunday if it’s still on the list. I have to get 15 points per day or I lose.

At first glance, it looks like Beeminder is basically a commitment contract: that it gets its force from promising to take your money if you lose. In my experience this seems very minor. I often forget how much money is riding on goals, and seem to keep the ones with no money on about as well as the others. So at least for me the threat of losing money isn’t what’s going on.

What is going on? I think Beeminder – especially the way I use it – actually does a nice job of combining a bunch of good principles of motivation. Here are some I hypothesize:

Concrete steps

In order to use Beeminder for a goal, you need to be clear on how you will quantify progress toward it. This means being explicit about the parts it is made of. You can’t just intend to read more, you have to intend to read one philosophy paper every day. You can’t just intend to do your taxes, you have to intend to finish one of five forms every week. You can’t just intend to ponder whether you’re doing the right thing with your life, you have to intend to spend twenty minutes per week thinking up alternatives. Making a goal concrete enough to quantify it destroys ugh fields and makes it easier to start. ‘What get’s measured gets done’ – just making a concrete metric salient makes it easier to work toward than a similar vague goal.

Small steps

To Beemind a goal, you need to divide it into many small parts, so you can track progress. ‘Finish making my presentation’ might be explicit enough to measure, but the measure will be zero for a long time, then one. Breaking goals up into small steps has nice side effects. It removes ugh fields, induces near mode, makes success likely at any particular step. In Luke Muehlhauser’s terminology, it increases ‘expectancy’ and allows ‘success spirals’*. Trading long term goals for short term ones also avoids the kind of delay that might make it easy to succumb to procrastination.

Don’t break the chain 

Otherwise known as the Seinfeld hack. This might be the main thing that motivates me to keep my Beeminder goals, in the place of the money. Imagine you are skipping rope. You have made it to 70 skips. It was kind of hard, but you’re not so exhausted that you have to stop. You probably feel more compelled to keep going and make it to 80 than you did when you started. In general, once you have successfully done something a string of times, doing it again seems more desirable. Perhaps this is particular to OCD kinds of people, but a Google search suggests many find it useful.

Beeminder is a nicely flexible implementation of this, because the chain is a bit removed from what you are doing. You only have to maintain an average, so you can work extra one day to slack off the next. This doesn’t seem to undermine the motivational effect.

Hard lines in middle grounds

Firm commitments are naturally made to extremes. This is partly due to principled moral stances, which tend to be both extreme and firm. But that’s not all that’s going on. It’s hard to manage a principle of eating 40% less meat. If people want to eat less meat, they either eat none at all, or however much they feel like pushed down in a vague fashion with some bad feelings. The middle of the meat eating spectrum is too slippery for a hard line – it’s hard to tell how much you eat and annoying to track it. ‘None’ is salient and verifiable. In other realms intermediate lines are required: your diet can’t cut eating to zero. So often diets are more vague; which makes them harder to keep.

Similarly, it’s easy to commit to doing something every day, or every Sunday, or every month. It’s harder to commit to do a thing 2.7 times per week on average, because it’s awkward to track or remember this ‘habit’.

Compromised positions are often more desirable than extremes, and desired frequencies are unlikely to match memorable periods. So it’s a pity that vague commitments are harder to keep than firm ones. Often people don’t make commitments at all, because the readily available firm ones are too extreme. This is a big loss.

Beeminder helps with making firm commitments to intermediate positions. Since you only ever need to notice if the slope of your data isn’t steep enough, any rate is as easy to use as a goal. You can commit to eating 40% less meat, you just have to estimate once what 40% is, then record any meat you eat. I’ve used Beeminder to journal on average five nights per week. This is better than every night or no night, but would otherwise be annoying to track.

A small threat to overcome tiny temptations

While working, there are various moments when it would be easier to stop than to continue, particularly if you mostly feel the costs and benefits available in the next second or so, and if you assume that you could start again shortly (related). It is in these moments that I tend to stop and get a drink, or look out of the window, or open my browser or whatnot.

Counting short blocks of continuous time working pretty much solves this problem for me. The rule is that if you stop at all the whole block doesn’t count. So at any given moment there might be a tiny short term benefit to stopping for a second, but there is a huge cost to it. In my case this seems to remove stopping as an option, in the same way that a hundred dollar price on a menu item removes it as an option without apparent expense of willpower.

I originally thought it would be good to measure the amount of work I got done, rather than time spent doing it. This is because I want to get work done, not waste time on it. But given that I am working, I strongly prefer to do good work, fast. So there’s not much need for an added incentive there. I just need an incentive to begin, and one to not stop when a particular moment makes stopping look tasty. In Luke’s terminology, this kills impulsiveness.

Less stress

The long term threat of failing to write an essay is converted into a short term pleasure of winning each night at Beeminder. I’m not sure why this seems like a pleasure, rather than a threat of losing, but it does to me. Probably because losing at Beeminder isn’t that unpleasant or shameful. And how could getting points or climbing a scale not seem like winning? (This is about value in Luke’s terms).

More accuracy

It’s harder to maintain planning fallacy, overconfidence, or expectation of perfection in the future, in light of detailed quantitative data, and a definite trend line.

Just the difference between ‘I should do that’, and ‘I should do that, so how much time will it take?… About two hours, so I guess it should get 20 points.. that probably won’t be enough to compel me to do it soon, but that’s ok, it’s not urgent’ seems to change the mindset to one more sensitive to reality.


In sum, I think Beeminder partly works well because it causes you to think of your goals in small, concrete parts which can easily be achieved. It also makes achieving the parts more satisfying, and strings them into an addictive chain of just the right challengingness. Finally it lends itself to experimentation with a wide range of measures of success, such as measuring time blocks or ‘points’, at arbitrary rates. The value from innovations there is probably substantial. So, averse as I am to giving lifestyle advice, if you’re curious about the psychology of motivation in humans, or if you want to improve your life a lot, you should probably take a look at Beeminder.

*you can also increase expectancy by measuring something like time rather than progress.

GD Star Rating
Tagged as: , ,

Dirty Air Kills

I’ve long been struck by how consistently different methods find large health harms from air pollution. Most people seem to think we no longer have an air pollution problem, because we mostly don’t see much air pollution. But the particles that are too small to see continue to cause great harm.

The US Federal EPA standard for air pollution in the form of particles of size 2.5 microns or smaller is an annual average of 15, and a 24 hour average of 35, micrograms per cubic centimeter. Many places are not in compliance with these standards (check your area here and here).

A 2009 paper in the New England Journal of Medicine estimated that decreasing this pollution number by 10 units on average increases lifespan by 0.61±0.20 years. A 2006 paper in the American Journal of Respiratory and Critical Care Medicine estimated that such a change would decrease mortality by about 15%, adding about two years of lifespan. (Quotes below.)

These are huge gains, which could be achieved at a modest expense, especially compared to the vast costs we pay for tiny health gains via medicine. More should be done.

Those promised quotes: Continue reading "Dirty Air Kills" »

GD Star Rating
Tagged as:

Shoo Freethinkers

Five years ago I wrote:

On average people who support odd ideas are less desirable as associates, and less discriminating in which ideas they endorse. If people only endorsed odd ideas when they had new information suggesting such ideas were promising, we should be eager to hear of such news, and eager to associate with such people. But in fact the main task faced by those with good news on odd ideas is to distinguish themselves from freethinkers who just pretend to have such news. Contrary to their self-image, undiscriminating freethinkers are our main obstacle to innovation. (more)

While giving a talk today on futarchy, I noticed how often freethinker fans are an obstacle to my innovation in particular. Mentally sloppy freethinkers tend to be attracted to radical proposals, just because such proposals are radical. They don’t focus much on the detailed arguments, but instead substitute simple arguments based on broad crude analogies, more suited to their style of thinking. And they usually make sure to insinuate that opposition to the idea is mainly from excess conformity or entrenched interests. Others hear such sloppy arguments, reject them, and then reject the idea as well.

For example, some say they like prediction markets because such things are markets, and all markets are good. This of course tempts others to reject them as based on knee-jerk free-market ideology. Some say they like prediction markets because they emphasize the wisdom of crowds, too long slighted by self-serving over-rated elite experts. Which elicits rejections from those who know just how often experts know better than crowds. Today someone even said futarchy was good because it is just like cost benefit analysis, which is obviously good.

Most big changes are bad ideas. So if a big change is a good idea, it must be because of some rather specific detailed reasons. When I make a radical proposal, I offer such specific detailed reasons in support, and those are the reasons I want skeptics to consider. For example, I argue for the information aggregation advantages of subsidized speculative markets, not for all possible advantages of all possible markets.

So when sloppy thinkers, eager to affirm their liberality by supporting radical proposals, latch on to my idea, and then substitute their own arguments based on vague analogies, they get in my way. Others see their support, and their sloppy thinking, and naturally want to distance themselves from the whole thing. Yes indeed, undiscriminating freethinkers are one of our main barriers to innovation.

GD Star Rating