Unfortunately, there are few prediction markets in the world, and fewer conditional ones, so it is hard to know how practical this institution may become. Thanks to an initiative by OB contributor Peter McCluskey, Intrade – the large, real-money prediction market – has added conditional claims based on the outcome of the U.S. Presidential elections. These may represent the first major case study of real-money conditional futures markets. So how are they doing? Here is some data and analysis.

Because of how the conditional claims are computed, we need to normalize the results by dividing the claim prices by the probability of a win for that party’s candidate. At the moment I looked (2008-07-28, 5:00 PM PDT), clicking on Politics and “2008 US Election”, and expanding “2008 Presidential Election Winner (Political Party)” I saw that the odds for a Democratic victory are 66.7-67.4. The claims are based on Democratic vs non-Democratic victories, so we will use the complementary odds of 32.6-33.3 for a non-Democratic win.

Then clicking on the left on “US Pres. Decisions” gives us the six conditional markets funded by Peter McCluskey.

Here are the claims, the prices, and the normalized values found by dividing by the corresponding Democratic and non-Democratic victory probabilities above. To compute the normalized values, I divided the lowest claim price by the highest party-victory price, and the highest claim price by the lowest party-victory price, giving the maximum range consistent with current trading prices. (Peter provides continually-updated information on current implied values as well.)

These prices imply that the market expects that under a non-Democratic administration, we will see substantially higher government debt, as well as much higher numbers of soldiers in Iraq.

In addition, there are two claims based on the chance that futures prices will move in the same direction as the Democratic party victory claim price, on election day:

These values are so close to the 50% mark that it appears that the markets do not expect any significant movement in oil prices or interest rates on election day, that can be attributed to developing information about which party will win. As critics have noted, this could be because they don’t see much effect of political parties on these values, or else because they expect that the election day results will be a foregone conclusion and there will be no surprises in that regard. (Or perhaps, that no significant information will leak out regarding election results during trading hours.)

Intrade has another set of conditional election claims under the heading of “Impact of Next Pres.”. Rather than party based, these claims are linked to specific candidates, meaning that at this point only the Obama and McCain claims have significant relevance. The claims measure economic growth, unemployment, crime, and control of the legislature, conditional on candidate victory – all interesting topics. But there is a huge difference in liquidity and trading volume for these claims versus the ones discussed above. These claims have bid-ask spreads that range from large to enormous; some have no bid or ask offers at all. And the trading volume is so far non-existent – apparently none of them has ever traded! These claims are essentially useless for shedding light on likely impacts of the presidential choice.

So what is the difference? The first set of claims is being funded and subsidized by Peter McCluskey, who has set up an automated market maker algorithm (based on an invention by Robin Hanson) that keeps the bid-ask spread within 2.5 points. When I first heard of the idea of such a bot, I thought it would have to be willing to lose infinite amounts of money, but actually this is not the case; losses are bounded. Nevertheless the potential costs to Peter are very significant, and he could be out almost $50,000 to fund his six claims.

Still, we can see the very real and practical benefits of this subsidy approach. The claims sponsored by Peter have relatively good liquidity and produce meaningful price comparisons. While not intrinsically more interesting than the other batch, Peter’s claims trade relatively frequently since there is always an attractive offer, while the others have never traded at all.

The lesson, then, is that conditional prediction markets can be informative, but that there is great importance in the detailed structure of the institutions. Peter’s subsidies have made the difference between a failed market and a successful one. At the same time, we see that the claims based on price movements on election day are not working as well as the others, so that idea may not be as promising (perhaps circumstances will change as election day approaches).

Another issue I would see as a potential trader is the difficulty of protecting my position against shifts in prices of two different markets. The value I am betting on is the ratio of two sets of prices, so I may want to change my holdings if either of those prices moves. It would be nice if one could post bids which are based on the ratio rather than on specific prices in the individual markets. It would also be nice to see the imputed values based on price ratios live and next to the market prices, rather than on another web page.

Even with these problems, it appears that the overall concept is a rather spectacular success that suggests that the concept has great promise. We all owe a debt of thanks to Peter for making such a substantial contribution towards showing the practicality of conditional futures markets.

You show up for an experiment and are told that you and a person arriving later will each have to do a different task on a computer. One job involves a fairly easy hunt through photos that will take just 10 minutes. The other task is a more tedious exercise in mental geometry that takes 45 minutes.

You get to decide how to divvy up the chores: either let a computer assign the tasks randomly, or make the assignments yourself. Either way, the other person will not know you had anything to do with the assignments.

Now, what is the fair way to divvy up the chores?

When the researchers posed this question in the abstract to people who were not involved in the tasks, everyone gave the same answer: It would be unfair to give yourself the easy job.

But when the researchers actually put another group of people in this situation, more than three-quarters of them took the easy job. Then, under subsequent questioning, they gave themselves high marks for acting fairly. The researchers call this moral hypocrisy because the people were absolving themselves of violating a widely held standard of fairness (even though they themselves hadn’t explicitly endorsed that standard beforehand).

I must admit that I too would probably assign myself the easy task. However I would hope that I would not be so hypocritical as to claim that I had behaved fairly when I did so. But of course, reading about the experiment is different from being part of it. Maybe I would have been just as hypocritical as the other subjects.

For me, the most interesting finding was that a simple intervention could eliminate the hypocrisy (although not the unfair action!):

[The researchers] brought more people into the lab and watched them selfishly assign themselves the easy task. Then, at the start of the subsequent questioning, some of these people were asked to memorize a list of numbers and retain it in their heads as they answered questions about the experiment and their actions.

That little bit of extra mental exertion was enough to eliminate hypocrisy. These people judged their own actions just as harshly as others did. Their brains were apparently too busy to rationalize their selfishness, so they fell back on their intuitive feelings about fairness.

This is an intriguing and (to me) counter-intuitive means for preventing at least one pervasive bias. It seems surprising because we normally think of our biases as being subconscious, in opposition to our conscious goals of clear thinking. In that sense, we might expect that distracting our conscious minds by memorizing numbers would actually increase the opportunity for bias to creep in. Yet in this case, we see the opposite. Apparently our subconscious evaluations of morality are more accurate and it is our conscious, volitional efforts which produce distortion.

Eliezer Yudkowski reported earlier on studies that showed the opposite effect, where distraction made people less accurate. It raises the question of what effect distraction would have on other forms of bias we have considered. I wonder if there is a rule of thumb for when one might productively use mental distraction in order to improve decision making?

Take a guess. Now, take another guess, different from the first one, and average them. According to research reported on in The Economist, averaging the 2nd guess improves accuracy by 6.5%. Better still, wait 3 weeks before taking a second guess. Averaging now improves accuracy by 16%. (Story found via Slashdot.)

Here is the full report from Pschological Science.  Some excerpts:

It is important that neither group knew they would be required to furnish a second guess, as this precluded subjects from misinterpreting their task as being to specify the two endpoints of a range.

That could make it a little tricky to do this on your own; you have to try to make your first guess as good as you can, and then start fresh for the second guess.

This benefit of averaging cannot be attributed to subjects’ finding more information between guesses, because second guesses were less accurate than first guesses

Hmmm, how can second guesses be less accurate than first guesses, yet averaging them is more accurate than either? I suppose it means that the two guesses tend (on average) to bracket the correct answer, with the second guess farther away than the first one. That means that your first instinct to improve your guess is more likely than not to be in the correct direction, but go too far. Knowing this might allow you to improve your guesses even more.

Oh, and as for the airports? According to the CIA World Factbook, there are 14,947 airports in the U.S., and 49,024 in the whole world, so 30% of the world’s airports are in the U.S. For the record, I guessed 25%, and then 15%, so averaging didn’t help me. But in general this might be a useful trick to easily improve guesses.

A somewhat more elaborate alternative is Electoral Vote Predictor, which uses polls rather than prediction markets. However, they are almost always the same, to tell the truth. I followed this site closely in the 2004 election. Site operator Andrew Tanenbaum includes a frank discussion of the accuracy of the 2004 predictions.

Robin Hanson argued earlier that closely following the news is not that great an idea, and I have to agree. It is something of a bad habit, but I find myself doing it anyway. We have this instinct that choosing our Leader is as important to our lives as it was when we were a tribe of two dozen, and that we have similar influence over the result. Following elections and participating in politics activates these vestigial tribal instincts in much the same was as sports, with similarly futile results. It can be seen as another form of “information porn“, exciting and titillating but ultimately unproductive.

I wish they used a different algorithm in the ElectoralMap site. It appears that the methodology is to take the individual state Intrade predictions, and assign all the electoral votes for a state to the party that has a 55% or greater predicted probability of carrying the state. States with less than a 55% probability for either party are shown as undecided. What I would like to see instead is to multiply the number of electoral votes for a state times the probability that the party will carry that state, and sum those results per party. This way, states that are seen as solidly preferring one party will be given greater weight towards that party than those which are more uncertain.

Then to make it even more useful, run a Monte Carlo simulation where each state is randomly assigned to one party or the other, with the odds taken from the prediction market, to create a simulated election result. Repeat this many times to calculate the mean and standard deviation for possible outcomes. In this way, confidence intervals can be given for the predictions. It’s possible that we might see the predicted electoral votes stay roughly the same for some time, but the confidence interval might narrow or widen. I don’t know if there is some statistical trick to compute the C.I. without actually running simulated elections, just from the distribution of probability estimates, but if so, that would be even better. My suspicion is that realistic confidence intervals are going to show that the election will be too close to call and could easily go either way – at least, that would have been the case in the last few elections.

Why not do this programming myself? Well, I might have the ability to do that. Intrade has an API, but it’s not clear if it allows retrieving the necessary data. I wouldn’t know how to display it on a colorful map, but the math should be simple enough, if you can retrieve the market prices, so producing a textual result would be straightforward. Maybe I will try to work on it in my spare time.

We don’t often post these kinds of pointer articles here, although they dominate many blogs. In that earlier posting about the news, Robin wrote, “I avoid posts that should not be nearly as interesting a year before or after.” This one does not really qualify, although there may be some historical value to be able to go back and look at how market sentiment evolved over time.

The starting point is Harold Hotelling’s 1931 model of the economics of nonrenewable resources. Hotelling showed that, based on very simple and general assumptions, nonrenewable resources like oil should follow a predictable price path. The price should increase gradually, year after year, with the rate of increase equal to prevailing interest rates. This price increase is likely to suppress demand and lead to gradually reduced production levels. So we see prices go up, demand and production go down, until the nonrenewable resource is exhausted and completely consumed.

Here’s a figure stolen from a website which stole it from somewhere else:

This is a rather complex diagram. The upper right quadrant depicts price on the vertical axis rising exponentially as time moves to the right. Then the upper left quadrant is a simple demand model which shows demand level on the horizontal axis decreasing as price rices on the vertical axis. We then move to the lower left quadrant which shows production level (which equals demand level, in equilibrium) on the horizontal axis decreasing as time on the vertical axis moves downward.

A couple of interesting points are, first, that prices become much higher than the costs to produce the resource, the opposite situation from competitive markets in other kinds of commodities. Yet resource owners restrain production without the need for a cartel or any coordination. They are motivated solely by their own individual profits. And further, this production schedule actually maximizes the net value to society of the renewable resource. An enlightened dictator would impose a decreasing production schedule in exactly the same form that Hotelling showed happens automatically in a free market.

However one of the most surprising aspects is that despite its generality and seeming soundness, the Hotelling model has universally failed to accurately predict the behavior of markets in nonrenewable resources. Oil has been produced for over a century, and rather than a gradual and predictable increase in price, its price was generally stable for most of the 20th century, with various ups and downs. Then in the past few years and months its price has shot up far faster than the rate that Hotelling would predict, the rate of interest:

One possible explanation of this long-term failure is that in the past, oil was treated more like a renewable resource; new oil was reliably being discovered, as quickly as it was being consumed. However this has stopped being true for twenty years now, and we are finally seeing a realization that oil is very much a finite resource. It is possible that we may be witnessing a transition to an era where the Hotelling model dominates. In that case, the current rapid price increase can be interpreted as a transition from a renewable-resource price to a nonrenewable-resource price.

Where, then, will this price transition end up? Although the Hotelling prediction of the yearly change in prices is robust and simple, deriving an initial price level from his model is quite difficult and uncertain. The basic idea is to postulate a so-called “backstop” price for the commodity, the price at which there is an adequate substitute, where demand for the commodity will drop to zero. Then the initial price level should be such that as the price rises gradually, at the rate of interest, and demand drops due to the price increase, we reach the backstop price at exactly the same time that we run out of the commodity; in the case of oil, as the last economically recoverable drop of oil is coming out of the ground.

I’ll illustrate with a sample scenario, but it will be clear that many alternative assumptions are possible. Let us assume that oil uses can be replaced, in the fullness of time, by solar-powered electricity. A barrel of oil has 5,800,000 BTUs of energy, which is equivalent to 1700 kiloWatt-hours. Presently, solar electricity is available for about $0.30 per kWh in favorable locations such as the American southwestern deserts. However the effective price will be higher in less sunny locales; for example, in Germany the price is $0.50 per kWh. Depending on distribution scenarios, solar electricity may have to be produced relatively locally, meaning less favorable conditions, so we will adopt the higher figure as a starting point. Today, then, we could replace oil for 1700 * $0.50 = $850 per barrel using solar energy, if we had the infrastructure in place to run transportation on batteries and replace other uses of oil.

Solar electricity is likely to fall in price due to technological improvements, although demand increases may moderate that fall. Installed solar prices have not decreased for the past 5 years, for example, as demand increases have more than kept up with production improvements due to economies of scale and improved technology.

As far as the Hotelling price increase, the rate of interest in constant dollar terms is generally quoted as being in the 3-5% range. Let us use a value of 4%, which implies a doubling time of 18 years. We will see oil rice gradually in price, as solar electricity gradually falls, and when they meet, we will stop using oil. All that time we are converting our infrastructure to rely on electricity rather than oil. At the point where the prices meet, we should have used up our oil.

So how much oil is there? This is also an area of controversy, largely due to assumptions about how much more oil will become economically recoverable because of higher prices, and also the possibility of future oil discoveries in relatively unexplored areas such as the arctic sea floor. My personal opinion is that the more pessimistic estimates are more likely to come true, for two reasons. In the first place, the rate of new discoveries over the past 10-15 years seems lower than would have been predicted. Second, many of the estimates of economically recoverable oil make an obvious mistake, in that they do not take into consideration the increase in expenses due to higher energy costs, when estimating what price would allow producing certain resources. Therefore, for this rough estimate I will adopt a more pessimistic scenario, corresponding to the USGS 95%-likelihood low bound.

This graph from the US EIA illustrates a number of possibilities:

The blue lines represent the most pessimistic scenario, an ultimate recovery of 2.2 trilliion barrels. The green lines reflect various demand scenarios. These are uninformed by the Hotelling model, so I would use the horizontal green line to approximate demand which is suppressed by constantly rising prices, until we are almost out of oil, at which time a blue line falls from this horizontal green line to zero. This zero point is only approached asymptotically in this graph, but a ballpark figure for substantially reduced production levels would be the decade of the 2060s, a bit more than 50 years in the future.

Going back to our oil price doubling time of 18 years, this is about 3 doublings, implying a price change from today to the endpoint by a factor of 8. With today’s backstop solar price of $850 to replace the energy in a barrel of oil, this corresponds to $106 per barrel as an estimate of what today’s oil price should be. To the extent that solar electricity falls in price over the next 50 years, the backstop price will be lower. If it falls by a factor of 2, the Hotelling oil price today would be only about $50.

These considerations suggest that the recent oil price of about $130/barrel is unsustainable in the medium term (over the next few years). If we maintained this price, oil demand would be over-suppressed, too much effort would be devoted to conversion to solar and other alternatives, and we would find ourselves in a position where we could stop using oil long before we needed to. This would hurt oil producers in the long run, as they would find demand dropping for their product while they still had plenty of easily recoverable oil available. Worse, our resources in this critical time would be mis-allocated and the net wealth available to future generations would be reduced.

However it’s also worth noting that we were at an arguably reasonable price level only a few months ago. The recent price spike may turn out to be a short-term bubble and we might well see a return to prices in the $50-100 range within the next year or two. That would put us at a sustainable price level that would allow for steady price increases, doubling every 18 years or so, as alternatives ramp up and are ready to fully replace oil in about 50 years, just as we run out. It will be interesting to see how closely actual events approximate this theoretical scenario.

It’s a provocative statement, but hopefully readers of this blog will have been exposed enough to the reasoning to accept it provisionally. Even Eliezer’s recent explanations of his various disagreements largely come down to making cases for why his disputants should agree with him, not for why they should all continue to disagree.

When two people disagree, and they come together to try to reach agreement, they have much to gain. First, some of their disagreement may be based on different information. By explaining the basis for their beliefs and sharing their data, each can improve the quality of his own estimates. Second, they probably have different biases affecting their reasoning. Discussion will illuminate those biases and help to cancel them out. Third, not being perfect Bayesians, they are computationally limited, and one or the other is likely to have superior reasoning, algorithms and heuristics. They can both aim to incorporate the results from the best quality reasoning available to the two of them.

Given these advantages, why then are people seemingly so reluctant to reach agreement? I think it can be easily explained in terms of human social status. All too often we view disagreement and debate as a contest, with a winner and a loser. The one who convinces the other of the correctness of his position wins. If they both reach an intermediate conclusion, this is seen as a tie, perhaps with an edge to one side or the other.

This viewpoint makes pragmatic sense in human society. Someone who is frequently right and convinces others of that fact is likely to be correct on other issues. He will be more trusted and respected as a source of advice and wisdom. Someone who often turns out to be wrong on disagreements is not going to be very well trusted in general. So people who win disagreements gain status, respect, power and influence, all of which lead to improved quality of life.

This effect makes people very reluctant to admit that they are wrong and to adopt the other person’s side in a disagreement. In fact, we often see a sort of bargaining process, in which adopting certain aspects of one’s opponent’s views leads to demands that the other side adopt something of one’s own views in return. Agreement is seen as very much a process of compromise and negotiation, rather than an objective search for truth.

Of course, OB participants are far above such mundane considerations as power and influence. We are clear minded seekers after truth, right? Right? Okay, maybe we do retain some vestiges of these natural human instincts. What can we do to overcome them and approach disagreement from a more Bayesian perspective?

One idea is to practice overcoming disagreement first on issues that are relatively easy. As with other areas of life, we do best by starting with easy problems before moving on to harder ones. Overcoming disagreement on matters where you have no emotional stake, no firm commitments, should be feasible. Yet often we do find ourselves disagreeing with others on quite trivial matters. These are good topics for practice.

It will be important that your practice partner is aware of the basic principles of rationality and the implications for how honest and respectful people interact when they disagree. You both should understand that “agreeing to disagree” is a sign of mutual disrespect and contempt. Rationality imposes a strong imperative to reach agreement and this must drive your interaction.

Moving beyond trivial matters, there are other strategies we can employ to make it easier to reach agreement. One is to avoid prematurely staking out a strong public position. Once you are committed publicly to a view, and your disagreement partner is likewise committed to an opposing position, it will be hard to avoid the winner/loser paradigm.

I tend to believe that an most issues where disputes are common, for most people, the evidence is really quite ambiguous as to which view is correct (as suggested by the mere fact that different people have reached different conclusions). The best position to take is a weak one, to hold views provisionally and to be open to persuasion. Adopting this as one’s public stance can actually improve status in many circumstances, since we all claim to admire those who have open minds. This kind of positioning can reduce the loss of status from changing your mind in a dispute, making it easier to reach agreement and gain the benefits of improved accuracy.

The final strategy I will suggest is the hardest, which is to renounce the social game and accept the possible lowering of status in disagreement, achieving a zen-like equanimity in the face of social disaster. This will not be easy but frankly, I suspect that many OB readers are already somewhat alienated from popular human social mores. Taking another step and consciously accepting the loss of status from being shown up as an intellectual inferior should be a reachable goal for many of us.

To move towards this ideal, consider taking actions which may accustom you to similar losses of status as you would experience from changing your mind in a disagreement. These might include making frequent, falsifiable predictions, many of which will inevitably turn out to be wrong; commenting on issues even where you are not too knowledgeable; sharing your speculations and thoughts even when you expect that they will lead to criticism. Air the dirty laundry of your mind, expose your ideas with all their unpolished flaws. In a world where most people build up a false front and do their best to hide their weaknesses, these honest actions can paradoxically make you seem mentally inferior. Such exercises can hopefully prepare you emotionally for being able to honestly report your changes of mind in disagreements.

Now it might be argued that this strategy could backfire, by hurting your reputation so that in the disagreement, your views will not be given appropriate weight. Indeed, this approach does depend on both parties being able to rationally evaluate and weigh their respective strengths, insights and quality of information. Ideally, then, both participants in the disagreement will be practiced at status reduction exercises, preparing them both to achieve maximum gains from overcoming their disagreement.

When this condition is not met, overcoming disagreement may not be possible in practice. Still, a practitioner of these measures will be better positioned to improve his accuracy as a result of the attempt at agreement, since he will be less bound to his previous position. He can still hope to gain many of the benefits from overcoming disagreement, making his efforts worthwhile in the end.

In classic group-decision experiments like “guess how many beans in the jar”, you get less accurate answers if people call out their guesses one after the other, because they are revealing their adjusted beliefs, that take into account the social consensus (perhaps without realizing it). If people write their answers down, we get Rolf’s kind of beliefs, uninfluenced by the consensus view, and those have been shown to be more accurate on average.

So Rolf’s point is very relevant about the lack of terminology. Devil’s Advocacy is about as close as I can come, but that doesn’t capture it. What do you suggest would be a good way to describe these kinds of beliefs? Once more people start making a conscious distinction between the two modes of believing, how should we talk about it?

Robin posted earlier about a NYT Magazine article on conformity. I was able to find an online copy of the scientific paper here: http://psr.sagepub.com/cgi/reprint/10/1/2.

The synopsis from the NYT is not complete. Of the 12 times that people were challenged to disagree with the social consensus, the most popular choice was to agree 0 times. 25% of the subjects did this. The second most common was to agree 3 times, done by 14%. Third most common was agreeing 8 times, 11%. Only 5% went along with the crowd all 12 times.

I think it’s quite significant that 25% of subjects never went along with the crowd and stuck to their own perceptions. In total, only 32% of the answers were wrong.

I’m not sure I follow Robin’s comments on this. It seems to me that this re-interpretation of the classic experiment suggests that people are not as conformist as generally thought. That would mean that we do more than merely give lip service to celebrating independence, that culturally we are quite effective at following the ideal of independent thinking.

The key question is, what is the right thing to do here? Should one conform when presented with 8 people denying the evidence of one’s own senses? I argue that it is the right thing to do.

Now of course, if you know you’re in a psychological experiment, maybe you can’t help but be suspicious that something fishy is going on. But in general, in real life, if 8 people come in and tell you that your perceptions are completely wrong, you should take it very seriously. I imagine that in the history of the world, in the great majority of such situations, the 8 were right and the one was wrong. As an example that some may be familiar with, if a bunch of friends come in and tell you you’re drinking too much, while your perception is that you can easily handle the alcohol, you should probably listen to them.

I would suggest that conformity is the right thing to do in these situations, and to that extent I am rather dismayed that the subjects were as non-conformist as this data shows.

My wife used to take Fosamax, so this was a topic of interest to me. To learn more, I did a search on scholar.google.com for “fosamax hip-fracture”. It looks to me like most of the hits are pretty favorable to the drug, but how much does that really prove?

From the 1st hit: “Therefore, for postmenopausal women with low bone mass, alendronate is well tolerated and produces significant, progressive increases in BMD at the lumbar spine and hip in addition to significant reduction in the risk of nonvertebral fracture.” From the 3rd hit, a recent meta-analysis: “We conclude that therapy with alendronate is associated with significant and clinically important reductions in the incidence of hip fracture in women with postmenopausal osteoporosis. The overall reduction is consistent among different patient populations.” From the 5th hit: “These data demonstrate that treatment with alendronate reduces the risk of multiple symptomatic fractures during a treatment period averaging 4.3 years. The reductions were consistent across prespecified subgroups. This effect is evident early in treatment and is sustained.”

The issue that bothers me is, how are non-specialists to know what is true? What is the best methodology to follow for a lay person to reach the truth? In this case, should he go to the primary sources, study medical textbooks to learn the details of bone dynamics, read all the recent research to get up to date, look at results from the various drug studies, learn about the pharmaceutical industry and the incentives governing research and publication, weigh all this evidence and come to his own conclusion? That would be a big job. And it is not at all clear that this method is likely to succeed, given that there is substantial disagreement among people who have in fact performed all these steps.

Or should he do as I did, a quick perusal of a sampling of results, reading through a few abstracts to find key sentences in the conclusions, to get a sense of the overall consensus in the field?

Or should he rely on the various exposes and/or rebuttals which float about on the web, perhaps focusing on those which best fit into his individual prejudices, preconceptions and ideological biases? That seems to be what most people do.

For a lot of “big” issues it doesn’t really matter; the individual’s opinion has no effect. But for questions of personal health, nutrition and similar issues, this kind of information could make a big difference in many individuals’ lives. Nobody wants a hip fracture in old age, but OTOH they don’t want their jaw to dissolve or whatever other horrific side effect may manifest. Our lack of a social consensus on a good mechanism for approximating the truth on controversial issues seems to me to be a major stumbling block in improving health and welfare.

de gustibus non disputandum
(there’s no accounting for taste)

Can Eliezer’s reasoning about lotteries be extended to something like sexuality? The goal of our sex drives is reproduction of the fittest members of the species. But today we have interrupted this process so that most sexuality does not lead to reproduction. The underlying logic of the drive is no longer functional. Does this mean that our sex desires are irrational, and that we should consider sex to be as pernicious and harmful as lotteries?

Another example is our love of sweet and fatty foods, which are harmful to our health in today’s environment. The fundamental reasons we seek these foods no longer apply today. In a sense, this kind of hunger is irrational, in that satisfying the drive is actually harmful to us.

Where do we draw the line in judging whether our desires are rational?

Most of our discussions of bias here relate to beliefs about truth. A bias gives us an incorrect view of the truth, and we seek to overcome the bias so that we can see the truth more clearly.

But there is a larger and more inclusive sense in which we can look deeper for biases in our motivations and goals. This requires us to adopt a model where different parts of our minds have their own individual goals, which may be in conflict. One part (or more parts) of our minds is the talking part, and that is who is writing here. Our discussions are primarily among the talking parts. It somewhat goes without saying that the goals we discuss as most important are the goals of our talking parts.

From this perspective, one part of our minds, the talking part, can judge the goals and methods used by other parts, by its own standards. Our talking parts favor logic and rationality. They want to see that the goals of a system are effectively being achieved by the means it employs. And by this standard, other parts of our minds fall short. Our sexuality does not achieve its goal of reproduction. Our hunger does not achieve its goal of health. Our love of risk does not achieve its goal of proportional reward.

We can identify these failures as biases and perhaps seek to overcome them. It is likely to be more difficult than mere failures to perceive truth, via logical illusions and constrained rationality. Our drives are deep and powerful, and even if they are ultimately pointless that does not take away their power. But I would suggest that it does make sense to hold our evolutionary drives up for analysis and judgment at least in the terms proposed here, whether they achieve their goals in today’s environment. If not, we should try to overcome these ultimately pointless or even harmful urges.