This is the first part of a mini-sequence of posts on zero-sum bias and the role that it plays in our world today.

One of the most pernicious of all human biases is zero-sum bias. A situation involving a collection of entities is zero-sum if one entity's gain is another's loss, whereas a situation is positive-sum if the entities involved can each achieve the best possible outcome by cooperating with one another. Zero-sum bias is the tendency to systematically assume that positive-sum situations are zero-sum situations. This bias is arguably the major obstacle to a Pareto-efficient society. As such, it's very important that we work to overcome this bias (both in ourselves and in broader society).

Here I'll place this bias in context and speculate on its origin.

Where this bias comes from

It's always a little risky to engage in speculation about human evolution. We know so little about our ancestral environment that our mental images of it might be totally wrong. Nevertheless, the predictions of evolutionary speculation sometimes agree with empirical results, so it's not to be dismissed entirely. Also, the human mind has an easier time comprehending and remembering information when the information is embedded in a narrative, so that speculative stories can play a useful cognitive role even when wrong.

Anatomically modern humans appear to have emerged 200,000 years ago. In the context of human history, economic growth is a relatively recent discovery, only beginning in earnest several thousand years ago. The idea that it was possible to create wealth was probably foreign to our ancestors. In The Bottom Billion, former director of Development Research at the World bank speculates on the motivation of rebels in the poorest and slowest growing countries in the world who start civil wars (despite the fact that there's a high chance of being killed as a rebel and the fact that civil wars are usually damaging to the countries involved)

[In the portion of the developing world outside of the poorest and slowest growing countries...] growth rates may not sound sensational, but they are without precedent in history. They imply that children in these countries will grow up to have lives dramatically different from those of their parents. Even when people are still poor, these societies can be suffused with hope: time is on their side...If low income and slow growth make a country prone to civil war, it is reasonable to want to know why. There could be many explanations. My guess is that it is at least in part because low income means poverty and low growth mans hopelessness. Young men, who are the recruits for rebel armies, come pretty cheap in an environment of hopeless poverty ...if the reality of daily existence is otherwise awful, the chances of success do not have to be very high to be alluring. Even a small chance of the good life as a successful rebel becomes worth taking, despite the high risk of death, because the prospect of death is not so much worse than the prospect of life in poverty.

Neither the developed world nor the countries that Collier has in mind are genuinely good proxies to our ancestral environment, but like the people in the countries that Collier has in mind, our ancestors lived in contexts in which growth of resources was not happening. In such a context, the way that people acquire more resources for themselves is by taking other people's resources away. The ancient humans who survived and reproduced most successfully were those who had an intuitive sense that one entity's gain of resources can only come at the price of another entity's loss of resources. Iterate this story over thousands of generations of humans and you get modern humans with genetic disposition toward zero-sum thinking. This is where we come from.

For nearly all modern humans, the utility of zero-sum bias has lapsed. We now have very abundant evidence that the pie can grow bigger and that win-win opportunities abound. Both as individuals and as representatives of groups, modern humans have a tendency to fight over existing resources when they could be doing just as well or better by creating new resources that benefit others. Modern humans have an unprecedented opportunity to create a world of lasting prosperity. We should do our best to make the most of this opportunity by overcoming zero-sum bias.

New Comment
164 comments, sorted by Click to highlight new comments since:
Some comments are truncated due to high volume. (⌘F to expand all)Change truncation settings

This post should present experiments that demonstrate this "zero-sum bias", or at least thought experiments that more clearly describe what counts as instances of "zero-sum bias". As it stands, you begin by presenting a vaguely described hypothetical problem, and proceed by hypothesizing about "its" origin in evolutionary psychology. You build on sand.

My posting is much more speculative than I would like and I wish I were on more solid ground. My reason for posting anyway is if this bias exists (as I strongly suspect based on my own experience and observations, but cannot produce rigorous scientific evidence for) then discussing it is extremely important. A quick Google search shows that very few people have explicitly discussed the possibility of zero-sum bias.
Robert Wright's Non-Zero is a good popularisation of the ideas, but I can't remember whether/how much the alleged bias per se is discussed.
Also, more thought experiments to clearly describe what counts as instances of "zero-sum bias" to follow in future posts :-).
If I recall correctly, one of the classic examples was the purely economic effect of employment discrimination. If the majority excludes some subgroup from economically valuable jobs, the bulk of the reduction in the subgroup's income isn't a gain to the majority - it is just a deadweight loss from forcing the subgroup to perform less valuable work. At least the economic (not status!) motivation for the majority to do this looks like a "zero-sum bias"
Your example only works if both majority and subgroup are otherwise equal (i.e. subgroup does same quality/amount of work for the same wage). In real life conflicts of this type, the subgroup will generally work for lower wages, thus lowering average wage and reducing income of the majority.
I suspect, but am not sure, that your "real world" assessment is biased in almost exactly the way multifoliaterose is hypothesizing that most people are biased. When someone is rationally willing/able to work for lower wages (assuming they aren't be forced into it by expensive systems of repression) it creates what economists call a comparative advantage which is an opportunity for mutually beneficial cooperation. All the people who would have done the drudge work for wages that are low (but not that low) can switch to whatever their new comparative advantages are with more total wealth produced in aggregate, which can be traded back around. Academics have noticed for years that comparative advantage is a real phenomenon, but also that it is not widely understood and is frequently denied even when explained. This seems a likely candidate for the kind of bias that multifoliaterose is writing about. If I had any quibbles with the article, it would be that (1) the object level was ignored in favor of mere "topic introduction", (2) while drawing support from evolutionary hand waving rather than citation to strong experimental evidence, with (3) the assumption that it certainly is bias (rather than a reasonably accurate model of the world). I would have liked to have read about the subject itself and thereby learned something, rather than reading about the tragedy of the developing world and why evolutionary hand waving is valuable. A good place to look for grounded material on the subject might be the literature that grew out of George M. Mason's classic work on "peasant culture". His 1965 paper Peasant Society and the Image of Limited Good included this summary in the introduction: I know that there exists almost 50 years of academic literature downstream of this statement, but I know little about its data, controversies, or leading authors. If someone was looking for academic results from which to borrow content (to popularize the material as "relevant to persona


Academics have noticed for years that comparative advantage is a real phenomenon, but also that it is not widely understood and is frequently denied even when explained.

That's true, but on the other hand, economists often present the theory of comparative advantage in a way that's either disingenuous or shows their own lack of deeper understanding of what it actually says. What they usually omit -- either out of ignorance or for ideological reasons -- is that the principle of comparative advantage is fully compatible with various realistic scenarios where great masses of people get completely screwed over by the emergence of trade involving a new party that enjoys absolute advantage over them.

This is true even when all the assumptions necessary to derive the principle of comparative advantage hold. However, to make things even worse, its derivation involves some highly questionable spherical-cow assumptions, and the question of what can happen when these are relaxed in various realistic ways is rarely asked. In particular, the fact of capital mobility can wreck the usual simple model of comparative advantage completely.

Now, you say:

All the people who would have

... (read more)
I'm in total agreement that there are "spherical cow" problems with comparative advantage, like the fact that there will be retooling costs if things change and that economies of scale are important complicating factors... and... the physical world is just complicated... so yes to that :-) But unless I'm mistaken, you're going overboard by pairing "more realistic assumptions" with outcomes that are bleak to the point of absurdity. Horses were property. They didn't have property rights over anything, not even property rights over their own bodies to prevent a predatory species (humans) from using their body parts in more efficient ways when it suited us. Horses didn't have the ability to negotiate or trade so they are not the kind of entities which are capable of personally leveraging comparative advantage. Moreover, they don't have a deep and generic capacity to learn new skills the way humans do, so their "economic function" was fixed. Finally, cars aren't economic agents either. The horse example didn't add real world complexity to the standard Ricardian examples, it subtracted complexity and filled it in with the specter of people being analogically "carted off to the glue factory"! Where is the dispassionate reason? Where is the evidence? This seems more like predator-prey ecological modeling than economics >.< Based on experience, to make a meta claim like "something people usually think is X is actually not-X" you need to be able to think clearly about the object level and then think clearly about the mechanisms by which people normally understand the object level. You must defend "not-X" while simultaneously explaining "X and its incorrect justifications". The best way I know to do this sort of thing is to talk about the real world in excruciating detail and provide links for the education of the audience and to allow verification of facts reported by third parties. The "incorrect justifications" normally fall out pretty clearly once the object level is


But unless I'm mistaken, you're going overboard by pairing "more realistic assumptions" with outcomes that are bleak to the point of absurdity.

The point is that unlike what one commonly hears from the proponents of free trade, including many economists, comparative advantage by itself does not prove that everyone, or even a great majority, will end up better off -- since it says absolutely nothing about how the wealth produced in the new equilibrium will be distributed. Moreover, when the spherical cow assumptions are relaxed in some arguably realistic ways, not even the conclusion that the total produced wealth will increase is certain.

Mind you, I am not making an argument against free trade here, but merely pointing out that the theory of comparative advantage is typically used as a nearly thought-free shibboleth, not a sound argument. It simply cannot prove what its fans commonly claim it does. To make a valid argument for a free-trade position on any realistic economic issue, much more is necessary.

That essay by Krugman you linked is a typical example. It sounds as if a mathematician complained that people are incapable of grasping the Pythagorean the... (read more)

It took me about a week to find the minutes to read this and process it, but once I was done, it felt like one of those cases where there was some new and relevant things in it, but the new stuff wasn't relevant and the relevant stuff wasn't new. To be clear, all I'm saying is "there's is something to comparative advantage worth learning from which seems to bear out the more general 'zero sum bias' idea" and your position seems to be a much stronger (and due to over-reaching, false) claim that comparative advantage is some kind of silly intellectual fashion among pointy headed intellectuals that should never even be mentioned without heaping portions of warning and quibble on the side. As though a little dose would be poisonous rather than give people a taste for learning more, and the meal wouldn't be that nutritious even by the end. If this is the ground you're defending, and that's the ground I'm defending, I really think you're simply wrong. The core insight here is not whether free trade among many many people is always pareto efficient for every single member of the economy, but whether 2 individuals can gain via comparative advantage (which obviously they frequently can) and then whether N people can institute free trade among themselves in a way that is Kaldor-Hicks efficient. If something benefits 20 people a quite a bit, and it benefits them more than it hurts one person who wants to veto all of their ability to trade with each other (rather than trade with the one person), then I say that the one person should look at themselves in the mirror and feel guilty. They should train for a new job (and maybe should be given a kickback from the profits for retraining) and bring on the efficiency! :-) It is obvious that sellers of anything would prefer that they have no competition in order to get nice fat profit margins. Someone competing with them will be bad for that seller but will almost always be good for the customers. If you give someone the political
JenniferRM: I actually don't see how these claims are contradictory, and in fact, I'd say they are both true. Yes, comparative advantage is a non-trivial insight that has something useful to say about certain situations that occur in the real world. However, at the same time, it is used in an entirely wrong-headed way by many intellectuals, as a supposedly conclusive argument for things that it simply does not imply. At worst, and sadly quite often, it is thrown around as an entirely empty-headed ideological shibboleth. The typical mention of comparative advantage you'll see in practice, even by economists, barely rises above the level of "Brawndo's got electrolytes!" After all, what does the principle of comparative advantage say? It's equivalent to the simple mathematical observation that if a1, a2, b1, and b2 are positive real numbers such that a1/a2 < b1/b2, then for any positive d1,d2 such that b2/a2 < d1/d2 < b1/a1, we'll have d2b1 > d1a1 and d1a2 > d2b2, and this is also true in the special case when a1 > b1 and a2 > b2. (In the standard Ricardian example, the a's and b's are the coefficients of proportion between labor and production for each good in each country, under the assumption that production is a linear function of labor put into it, and d's are the amounts of labor that shift between the goods in each country when trade emerges.) Now, as I said, there really is some non-trivial insight to be gathered here. But if you believe that this simple piece of math is enough to model what actually happens when trade is liberalized or some other bearer of absolute advantage appears on the market, always and in all possible circumstances, and if you don't see various critical unrealistic assumptions and all the numerous relevant variables that don't even get considered by the model, then with all due respect, I can only conclude that you haven't thought about it much. Imagine an engineer designing a house, and using the Pythagorean theorem several times in

I think this might better be framed as a "relativity bias", or perhaps the "relativity heuristic"

We don't assume that someone else winning means we have to lose, we just estimate things based on relative rather than absolute reference level. The status game would be one obvious example of this, and Yvain wrote a great post on how judgments based on relativity power thought processes all the way down to the level of perception.

You can also see it in the somewhat famous studies that show someone will drive across town to save 5 bucks on a 10 dollar calculator, but not a 100 dollar jacket, the fact that we judge gifts based on their expense relative to their class rather than their absolute expense, and probably about 100 other studies that I'm forgetting at the moment.

I like this remark. Quite possibly the "relativity heuristic" explains everything that I have in mind when I refer to "zero-sum bias." I will have to think about this some more. Regardless, I think that it's worthwhile to isolate instances where the relativity heuristic pushes people in the direction of zero-sum thinking. From the point of view of people like myself who subscribe to utilitarian ethical principles, irrational zero-sum thinking is worse than some forms of irrational thinking, because when somebody engages in irrational zero-sum thinking, this has negative effects both on his or herself and on others, whereas some other forms of irrational thinking may not have systematic negative effects on others. Here when I say "irrational" I'm referring to failure to exhibit "instrumental rationality" as opposed to failure to exhibit "epistemic rationality."

I think you miss an important part of the ancestral environment that is a likely source of zero-sum bias. In the ancestral environment, material resources are far less important than status once you're not starving. The difference between the top and the bottom of the wealth distribution was much smaller than it is today. But holding more resources was still important - more resources means higher status (roughly anyway). And status is a zero sum game. So since all or most transactions essentially status transactions, zero-sum thinking was a good heuristic.

Of course, like your theory, this is highly speculative.

I agree that there's something to what you're saying. Your comment gives me occasion to make a point that I had been intending to make later. I anticipate that some people will argue that since status is zero-sum and since humans crave status, zero-sum bias still helps humans get what they want. To this imagined argument (not made by you), I would respond that to a large degree, humans don't crave status, they crave the feelings historically correlated with status, and achieving these feelings is not a zero-sum game. Similarly, historically the quest for reproductive resources was a zero sum game, but (a) this is no longer the case, and (b) we do not crave reproductive resources for their own sake, we crave the feelings historically correlated with attaining reproductive resources, and the quest for these feelings is not a zero-sum game.
The status game is not entirely zero-sum either.
Could you elaborate or point to a link about status being positive sum?
Sure. You can imagine two seemingly equal tribes. One with much more advanced status structure, where the chief is more revered, where there is a shaman with his own charisma or high status, where every member has it's own higher then zero place. A kind of Vanity Fair, but non the less. And we can play this game of status in a smaller groups as well. Vote me up, I'll vote you up and we will both gain the status. We will cut together a little bigger piece of karma cake for us. A nationalist leader may tell his people, that they are special. If they decide to believe him, the status of everybody will go up. At least they will think so, but it's all that counts in the status game, anyway.
So it IS zero, since there is less cake for everybody else. Helen Parr (to her son): "Everyone's special, Dash." Dash: "Which is another way of saying no one is." -- the Incredibles
Exactly. A person can only be high status by being higher-status than someone else; so one person's high status must lead to another person's low status. So status must be zero-sum.
5Eliezer Yudkowsky
I have no trouble visualizing a society composed mostly of people with high status, or a society composed mostly of people with low status, with very different sums of total status.
I think we might be using different definitions of status. So instead of status, I'll say that social power is zero-sum.
Many people may share social power, especially if they don't choose to wield it often or to the detriment of others. I suppose you'd say that you count them as having it in exact proportion to their tendency to actually use it, or in terms of the power they'd likely have if they chose to war against one another.
No, the point is if someone gains social power, someone else must lose that power. Sharing of power is fine in this framework - if you share power over the tribe, for example, then you don't have full power over the tribe. For one, you don't have the same kind of power over the individuals with whom you are sharing power.
You can gain social power that was previously held by natural randomness.
What if you construct more than one cake, then arrange distribution so everybody gets a bigger piece than somebody else on at least one cake. Thus, because of human tendency to emphasize what makes them feel good, people notice their privileged cake(s) and disregard their loss cake(s). A real-world equivalent would be the religious concept of poorness as a virtue.
I think the issue is whether to use "relative status" or "absolute status". For example using the karma example, it is not very important what the karma numbers are absolutely but what their relative value is. Thus a couple of friends voting each other up raise the average (+mode + whatever statistical marker one prefers). Thus while their absolute status rises the relative status of other people sinks. I think we may have different notions of status with me thinking of "relative inside a given group".
One word: Sex. A high status male would have access to more females. Remember, the point of evolution is not to survive for as long as possible, it is to f**k as much as possible (for a male).

Remember, the point of evolution is not to survive for as long as possible, it is to f**k as much as possible (for a male).

No, and it's not even to have as many offspring as possible. It's to have as many copies of his genes in future members of the species.

Consider to male proto-humans, Adam and Bob. Adam has sex with his sister, they have 6 children, but all die without reproducing.

Bob never has sex, but is a good uncle to his brothers' and sisters' kids, 4 more of which survive to reproduce than would have done without his interventions.

Which one was more effective at passing on his genes?

I"m not sure it's quite that simple-- some men are interested in raising children as well as conceiving them. And it's plausible that the best reproductive strategy is to attempt to have as much sex (with as many women?) as possible, but also to have a relationship with a woman (who one tries to get to be monogamous) and contribute to raising her children.
Do you know that the modal number of children fathered per male over our entire human history is zero?
I've heard something like this claimed before and it sounds plausible but I haven't seen a reference - can you point me to one?
Can't easily find it. Karma to the finder/refuter.

"Today’s human population is descended from twice as many women as men."

I think this difference is the single most underappreciated fact about gender. To get that kind of difference, you had to have something like, throughout the entire history of the human race, maybe 80% of women but only 40% of men reproduced.

From an address to the APA on gender differences delivered shortly after the Harvard/Summers business. Long and only tangentially related, but worth a full reading, IMHO.

It's not bulletproof in present context. The author doesn't cite primary sources and isn't an authority in the field. Still, given the extent and... energy of the backlash underway when it was delivered, I doubt that an uncorrected version would still be available from FSU's official web servers if an easy refutation was available.

According to Tierney, Baumeister's source is Wilder-Mobasher-Hammer (not gated, but also). An intermediate observation is that local mitochondrial Eves tend to be twice as old as local Y-Adams. The paper goes on to draw the conclusion that the effective sex ratio was 2:1. (which is not quite the same as a male mode of 0.) I would be pretty negative about this deduction, except for the last couple of sentences (the ones citing Shen et al and Hedrick) which claim that this is compatible with autosomal observations. I think you have wildly false beliefs about universities. Can you point to documents that were pushed off of university servers or corrected due to political pressure in the context of that backlash?
Er, shouldn't wrong papers be corrected or withdrawn, even in the absence of political pressure? Anyway, I'm not insinuating anything here. I'm just pointing out that controversial statements get aggressive fact-checking
thankyou, upvoted
Does that statistic (in so far as you remember) include males who died before growing old enough to implement a reproductive strategy?
I know that the mode for females was \2 in that same piece of data.
That's pretty suggestive, but male infants and children are less robust that females, so it's not conclusive (to me, anyway, because my background knowledge isn't good enough for even order-of-magnitude estimates of the numbers involved).
The average number is 2. For every man that had more, other men had fewer.
That's the mean average. I said modal, the mode of a discrete probability distribution being the value with the highest probability or frequency.
I am quite sure you are correct and the mode is 0. If males who died in childhood are included it almost certainly has to be correct. It must be correct if childhood mortality is 50% or greater.
I think that Roko's claim is that the mode among men who survive to adulthood is zero. See Roy Baumeister's Is There Anything Good About Men?
You seem to be describing a median average. If 1 man has 100000, while 999 men have 0, then the (mean) average is 100 per man. But for every man that has more than that average, there are 999 who had fewer. If someone says average without specifying, I take them to mean "mean average".
Jonathan, you are discribing a situation that never happened. The gender ratio is about 1:1. The average number of children is 2 because a different number than 2 results in numbers of humans we know did not happen. If one man had 100,000, then 49,999 men would have to have zero to make the averate equal to 2.
Thus "if".
No, I didn't, and I second the request for a cite. If true, what does it say about the best individual strategy? How sure are you that the evolutionary pressure is about the extreme long haul rather than just for a generation or three? If it is zero, any estimate of how much is the result of ingroup competition versus outgroup? To put it another way, how much should a reproduction-optimizing male invest in defending his group?
I don't understand this question
Is selection for having the most grandchildren, or is it for having the most descendants a millennium later? I'm not sure that's a real distinction, but I'm not sure that it isn't, either.
You should read the selfish gene. He has a chapter on this question. The answer is that the time period depends upon the longevity (in generations) of the piece of DNA that you take as your "basic unit of selection"; the gene isn't a precisely defined concept, rather Darwinian theory is parametric over the length of the basic unit of selection. In this case, it doesn't really matter.
For most purposes, selection should be on the long time scale. If the environment is uniform, long-term fitness should be the same as short-term fitness (though less random). If there are occasional catastrophes and population bottlenecks, then being adapted to them may be more important than being adapted to the usual environment. Even with a uniform environment, there may be a long tail to male fitness which is not easily observed in the short-term. Genghis Khan demonstrates that there is, at least occasionally, a long tail to human male reproductive success. If genetic factors were relevant and the opportunity arose reasonably often, then we should expect those genes to spread, even if they impede normal reproduction.
This is an important point, but can be stated directly, rather than bringing in the "ancestral environment" thing as justification. "Status" - or, more concretely, power over others - corresponds a great deal in our (and probably every actualizable) society to wealth. A society with big wealth differentials is going to be one where some members wield great power over others; it's not simply a question of differential consumption rights. I think this is basically motivates people to find wealth differentials unjust: as supporting evidence, consider that people don't consider happiness differentials that can't be translated into power differentials unjust - nobody thinks it unjust that Alice has a better native appreciation for classical music than Bob - and that the people who are most likely to consider huge wealth differentials just, libertarians, are the most likely to narrowly map power differentials to the exercise of physical force.

The flip side of this is when people assume that anything bad must have a counter-balancing "silver lining" that evens the overall outcome to exactly zero. This happens particularly often, for example, when I suggest that if it were possible to completely eradicate clinical depression then we should do so.

Yes, status quo bias also seems to have arisen from an evolutionary history of no/(very slow) growth.
I wonder whether either or both applies to the True Ending to "Three Worlds Collide".
I wonder whether either or both applies to the True Ending to "Three Worlds Collide".
I understand what you are saying, but depression happens to be a not good example. Depression is a necessary "feature" in the control system that physiology uses to modulate its hedonic state. During a near death physiological state, the state one can attain while running from a bear, where to be caught is certain death, physiology induces a state of euphoria. The near death physiological state has to be euphoric, so that organisms can willingly run themselves to death. This is the source of the euphoria of autoerotic asphyxiation (what killed David Carradine). That state has to be euphoric because all the “safeties” that prevent organisms from damaging themselves (pain, fatigue, extreme pain, extreme extreme pain) are turned off to allow even a very slim chance of escaping from a bear. If organisms could enter a euphoric state easily, they would, and would risk death uselessly. There must be an aversive state between “normal” and the euphoric near death state. I am pretty sure this aversive state is depression. I say that as someone who has been depressed most of my life, and has been on antidepressents just about half my life now, so I do understand depression. It isn't that depression is something good, but that an organism that can support a depressive state has superior survival characteristics to one that cannot. So evolution has configured all organisms to have the equivalent of a depressive state. I think the reason people imagine there is a “silver lining” to depression is because depressed people are easier to control and bullies feel less guilty when they bully someone into a depressed state because they fantasize there is a “silver lining” to the depression.
I'm not following this at all. Why isn't the aversive state "pain, fatigue, extreme pain, extreme extreme pain"? If you were running yourself to death, I'm pretty sure that's what you'd feel before the euphoria kicked in, not depression.
Blueberry, pain, fatigue, and depression signal different things. Pain signals local injury, fatigue signals not enough ATP in your muscles to do what ever it is you are trying to get your muscles to do, depression signals more of a global energy crisis, particularly in the brain. These things are to limit physical activity to reduce the damage that the physical activity will cause. The “energy crisis” of depression is in the brain, but the current experience of depression is more of a problem with malfunctioning of the control pathways that regulate brain ATP levels. Low ATP in the brain is associated with depression. There are some other examples, vascular depression which is common in the elderly, there is reduction in ATP levels in the brain coincident with reductions in brain perfusion. Fix the perfusion and you fix the depression (and vice-versa because they are coupled). Bipolar disorder has similar low ATP status. ATP status is extremely well regulated. Depression and allied mental disorders are disorders of that regulation (which is very poorly understood).
Alternate notion about depression: creatures capable of thinking about alternatives need a "sounds like a good idea-- do it" mechanism and a "sounds like a bad idea-- don't do it" mechanism. If the "sounds like a bad idea" mechanism is running too strongly, you get depression and/or akrasia. If the "sounds like a good idea" mechanism is running too strongly, you get mania.
NancyLebovitz, yes, almost exactly. My hypothesis of bipolar disorder is that the trade-off of normal-depression-mania is the normal-depression-euphoria of the near death metabolic state when one is running from a bear. I think what the euphoria does is more change the “discount rate” that organisms apply to their actions, immediate gratification vs long term gratification. When you might be dead in a few minutes (because the bear catches you), the time-value of risk-reward has different values. Any risk is worth taking if it might extend your life beyond the certain death of the bear catching you. I suspect that this is the same physiology behind the stimulant drugs of abuse; it triggers the same near-infinite discount rate, where continued good health a few days from now is worth nothing compared to the next injection from a shared needle which might have HIV in it. I think this is why deterrence has been shown to not work to deter drug abuse. The lives that addicts already live are worse then what you can impose on them as punishment. If the addict lifestyle doesn't deter drug abuse, certainly the relative cake-walk of a stay in prison isn't going to. The changing of the “discount rate” isn't under conscious control. It can't be because non-conscious human ancestors needed to do the equivalent calculations too (so it it from deep evolutionary time), and there isn't time (or cognitive capacity) to do those calculations consciously. I think this changing of the discount rate is part of the problem of zero-sum bias. To an addict, an injection of drug from an HIV infected needle is worth dying for. The shifting of the discount rate occurs on a continuum. Lesser amounts of stress shift it less. Trigger an infinite discount via stress-induced bullying, and victims will agree to anything to get the bullying to stop. I think this is a common ploy of some politicians.
Is there information about how it feels to run from deadly danger? I don't get the impression it's a euphoric experience, but evidence would be nice.
Extreme metabolic stress is euphoric, as in autoerotic asphyxiation and I think the runner's high. I think exercise addiction is addiction to the euphoria of extreme metabolic stress. Euphoria is a complex physiological state. I presume that for simple drugs to induce euphoria, they are actually triggering already existing pathways.
0Alex Flint
Thanks for taking the time to reply daedalus. Let's just say for the sake of argument that depression does have an evolutionary basis as you say (it would actually be more surprising if it didn't have any such basis). From my perspective this wouldn't automatically make it worth preserving if we were in the hypothetical position to eliminate it. That something is evolved, or has an evolutionary basis, does not automatically decide its goodness or badness -- evolution has given us many characteristics that we would rather change were we in a position to do so.
alexflint, I agree with you. Just because something has evolved (or may have evolved) is not a reason to continue it. I suspect that violence against women may have an evolved component (to reduce maternal death from cephalopelvic disproportion), but now cannot be justified for that because there are much better options (medical c-section). There may be other “features” of violence against pregnant women (epigenetic programming of the fetus to be more violent via the 'cycle of violence”), but I think the downsides of violence against women greatly outweighs any positives that may have existed during evolutionary time. I think if we could eliminate violence against women we should do it. What ever positives there might be for individuals, for society it is a net negative (in my opinion).
I'm not sure what your line of thought is about violence against women. Induced miscarriage?
Induced miscarriage, preterm birth, low birth weight. There is evidence that all of these are associated with violence against women, and also violence against women by her mate increases while she is pregnant. I looked and was unable to find any data suggesting any non-human males commit violence againt a female while she is pregnant with his fetus. Humans are unique among mammals for the degree of placental difficulties, birth dificulties and cephalopelvic disproportion they have.

Interesting topic, interesting quote. However, I am not sure we should regard the third-world rebels as victims of zero-sum bias -- they might understand that their country would be a better place if people worked together on building it up, but the chances of that happening soon, are rather low.

I'd say it's more of a tragedy-of-the-commons problem, where people are less interested in the negative sum of rebellion or piracy or ..., but are more interested in how much they personally gain. Which is of course not so much different from the developed world o... (read more)

Yes, my suggestion wasn't that third-world rebel's plight is primarily borne of zero-sum bias (although it may be in part), I was drawing a parallel between the situation of the third-world rebels and our ancestors.
What about the notion of a "negative sum" bias? I'm sure there is an appropriate technical term. The point being: what about the act of decreasing overall wealth as an act of personal enrichment, or of "relative improvement"? For example, if I am in competition for a resource like grazing land, if I kill my neighbor's cattle it leaves more resources for my wealth might increase even if overall wealth decreases. (this is something of a corollary to "Tragedy of the Commons") Alternately, if I am in competition for a resource like overall prestige, if I kill my neighbor's cattle, I have more relative to him, even if my wealth is unchanged and overall wealth is decreased. As a third issue, in relation specifically to civil war, there is "revenge bias". Do we have more evolutionary bias towards summation or wealth issues? Or is it more about "give me my endorphines"? When we feel slighted, we will destroy things because the act of it makes us feel better. Adrenaline trumps cortisol, is the evolutionary imperative here. Bad conditions, or bad actions, result in personal stress. Nothing relieves stress like violence, and revenge is especially sweet. All three of these, in the end, are about "how good do you feel as a result of an act". Extrapolating it to economic considerations might be going too far. The evolutionary imperative is more simple: the specific short term acts have physiological results, which are more tangible (in an evolutionary sense) than long term "wealth issues" (even while the long term wealth issue can reinforce the short term act as well). Heck, ever see a cat pee on a computer as an act of "rebellion"? Where's the summation of wealth there? The act of rebellion pre-exists the human condition.
Indeed, many of the of the 'zero sum games' we see in reality are actually more like negative sum games, as one man's gain is less than the other man's loss. However, I wouldn't say there's bias for that - in fact, it seems the negative sum is often not recognized. Or? Regarding the revenge bias, one reason sometimes mentioned for the worth some cultures (past, and somewhat less, present) cultures put on revenge comes from our pastoral past, where protecting your prestige, your honor would be of crucial value, so other will not steal your sheep that you can only loosely guard. The story then goes that the more closely connected a culture is connected to some pastoral past, the more value it will but on vendetta and revenge. Seems somewhat plausible, but hard to proof of course.
Well, the relevant fact about zero-sum games is that their sum is constant - as long as that holds, worrying about its exact sign seems not too important...
I'm not sure that the sum of "wealth" is constant, in a negative sum game. The willful destruction of someone else's wealth/resources does not result in a constant sum...that wealth is destroyed. So the point is, there are zero-sum games, and non-zero-sum-games,and the question is where does the bias lie, at any given time?
You seem to have misunderstood some terms here. A general negative-sum game does not have a constant sum, no. But the constancy here is not constancy of sum between before game and after game; if that holds, you have a zero-sum game. The constancy here is over all options the players can take. If that holds, you can subtract out that constant to obtain a zero-sum game with equivalent strategy. A person who assumes all of a certain class of games sum to "1", whatever that means, will have the same bias in his strategy as one who assumes that all such games sum to 0. The only difference is that he'd want to play more often.
I guess I need your analysis in a real world example, because I think we are talking too much about the "game" model. If I kill your cattle, or I salt your earth, what is the sum? What is the constant? What is the bias? My point is: the sum is negative, there is no constant, and the bias is towards "gratification". I don't think killing my competitors cattle comes from an inherent evolutionary-economic analysis...I think it comes from "doing this releases endorphines in my brain in the short term, I see his wealth destroyed and that seems good". The bias is simply "relative success"...I win by gaining more, by losing less, or by making him lose more than me. It's all very short term and emotional.
And going for relative sense makes sense when? In a zero-sum (or if you like, constant-sum) game. Though this may be getting away from the original statement?
What happens out there in the hurly-burly is not a "zero-sum" or "constant-sum" game. Specifically: it's not a "game" at all. Those games are distillations and models used for testing behavior. This tells us certain things about how people react and interact, but it doesn't tell the whole story. Going for relative makes sense when you can't take, you can't (necessarily) earn, you can't (in the short term) increase/generate. But you CAN win by destroying. You destroy your opponents resource, thereby "increasing" your wealth in a relative sense. It's why we bombed weapons plants during WW2, no? And to an extent, it's why they salted the earth....

You have neglected the negative-sum lose-lose situation being mislabled as win-lose.

War is the classic lose-lose situation that is mislabeled as win-lose. No one "wins" a war. After a war, everyone is worse off, just some are more worse off than others.

I think the problem is that the zero-point shifts, where if you survive a war, you feel like you have won something where in reality you just didn't lose your life.

At least in one point in history, it was possible to wage war as a strategic move, so that your country would gain more resources than you expended in war. This is probably not possible anymore, because of advances in weapon technology.
You can't really conquer another country and turn a profit on it these days, but it seems as though civil war can still pay: you can overthrow a dictator and install yourself in his place as the new dictator, and then allocate the spoils to yourself and your supporters. This seems to work best if there are exploitable natural resources (such as diamonds or oil) that you can gain control of.
Civil war or coup?
Well, coups work better, but civil war seems to pay off at least some of the time. (At least you can sometimes get de facto independence from a local dictator.)
Civil wars can be even better than a coup if you find external sponsorship.
Even in that case, the war as a whole is a negative sum. The sum of resources after a war is less than the sum of resources before a war. Net resources have been lost due to consumption in the war. If you discount the value of resources owned by your opponent in war to zero (while your opponent owns them), but not when you own them after the war you can come up with something positive, but I see that as an accounting gimmick.
Then you're missing the point. Historically, the kind of people who start wars of conquest tend to hold philosophies that value only their own in-group and its associated resources. When you don't value the happiness or prosperity of your victims it's quite easy to "win" a war. You just have to make sure the cost isn't so astronomical that it wipes out the benifits your in-group can reap by exploiting your victims over the next few generations.
It may still be negative-sum, but at least it's really win-lose instead of lose-lose.
This isn't entirely clear, depending on how you look at resources. It's possible that military training and research produced skills and expertise useful in peaceful arenas, yielding an overall gain. It's also possible that a conquering nation can exploit the resources of the conquered nation in a way that the conquered nation couldn't, yielding an overall gain of available resources. The mineral wealth of Afghanistan might be one real-world example; the resources were there, but undiscovered and unmined.
If you look at resources before and after the war with the same metric, then resources have been expended and lost. If you are using a different metric to measure resources before and after the war, then it is the metric that has changed, not the positive sum generation of resources. Once you start killing people (as is an inherent part of war), then any talk of gains and losses goes out the window unless you attach a specific value to specific human lives before and after and are willing to compare those lives lost with material resources. Since victims usually attach a higher value to their lives than do perpetrators, mutually agreeable values for the gains and losses can not be achieved. Compelling people to do something against their will (i.e. slavery) is a negative that can not be “balanced” by what ever positive things the slaves might generate. That is why slavery is wrong, no matter how “productive” the slave masters compel the slaves to be. You do not make slavery “less wrong” by compelling the slaves to be ever more productive. Once you privilege the values that you attach to things, then so long as you have gains, then you will perceive every interaction to be positive-sum because you have gained even if everyone else loses. I think that the original poster was meaning “zero-sum” in circumstances where all parties have equivalent knowledge, that a transaction with asymmetric information (such as where one party knows the flea-market painting is a rare masterpiece and worth millions and not the $10 sticker price) isn't in the same class of transactions. In other words the idea is that the transaction actually be positive sum and agreed to be positive sum before and after the transaction. A transaction with asymmetric information is more of (as I see it) a “gaming” the transaction and doesn't really achieve a true positive-sum. The Afghan war didn't generate resources, those resources were always there. Buying the $10 masterpiece didn't turn it into
OK, we'll stick to the same metric. Do you agree that human skills and abilities can be resources? Do you agree that the ability to exploit a resources is itself a resource? Suppose country X has a functioning democracy, court system, banking system, and corporations, while country Y has none of those things. If country X invades country Y and sets up a functioning government and economic system, this could well be positive-sum. Country Y didn't gain any material resources, but it now has a greater ability to exploit its natural resources as well as its human potential. This is a clear positive-sum situation. One study put a value of around $1.5 million on a human life. You're right that people will be biased when they try to value their own life, so we should probably disregard any self-assessed value. Of course we need to be able to value lives and trade them off against other resources; we do it all the time when we make policy or safety decisions. Have you read the articles on this site about utilitarianism and deontology? This sounds like a deontological position; I think most of us on this site would disagree. Not about slavery being wrong, but about why it's wrong: that the harm to humans outweighs the benefits.
Simply discounting self-bias in valuing a life doesn't give you a correct value. The opposite of self-biased is not unbiased. Human skills can be a positive resource. What ever skill are generated during a war, it does not take a war to generate those skills. Those skills could be developed in the absence of war. That those skills are not developed in the absence of war is not an argument I find persuasive that war has provided the benefit of the development of those skills. I don't consider that responses that people make to mitigate adverse circumstances can ever completely negate the adverse consequence. I think the idea that people have that a “silver lining” can completely mitigate an adverse event is part of the zero-sum bias the OP was talking about. Maybe if the war had not happened, then even better skills would have been developed and without all the damage the war brought. Part of the issue is that different events and consequences are to some extent orthogonal and can't be directly compared against each other. Part of that is that we can't know the actual consequences of paths not taken. Maybe one of the victims of WWII would have gone on to invent something that would have triggered a phase change in space-time and destroyed the whole universe.
I'm not sure it's even meaningful to put a dollar value on your own life. And, yes, we'd want to correct for the biases of hating the person in question, as well, which might lead you to undervalue him. I'm not claiming that war is the optimal outcome, only that it could be positive-sum. There may be other choices with an even larger sum. However, maybe the only way to exploit country Y's resources is with war; suppose all diplomatic attempts seem doomed to failure. Then, if the resources are valuable enough, war might be the best option (though we should also take into account how easy it is to underestimate the cost of war). No, the zero-sum bias consists of erroneously thinking that an adverse event is always mitigated by a silver lining. Adverse events are sometimes mitigated by a silver lining. For instance, if you would freely choose to get a papercut for $10, and someone gives you a papercut accidentally, couldn't they mitigate the adverse event by giving you $10?
I think the issue of lives in the context of "sums" is this: how many lives did "we" lose, compared to how many lives did "they" lose, in order to come to a conclusion of the conflict in and of itself. The sum is only self-referential....what happens afterwords is not relevant to the argument. e.g. in a $10 zero sum experiment, the "winner" leaves with $9 and goes and buys crack on the street. The "loser" takes his/her $1 and buys a winning lottery ticket. The long-term winning and losing after a war is not quantifiable, because there are no controls. Too many decisions, laws, random chance, weather events, could have taken things in one direction or another...who's to say?
It depends on what one assumes the motives for war are. If they are economic then I think a case can be made everyone ends up worse off. But if power is at stake, then war can indeed leave the nominal victor better off (from the perspective of motive). By the way, attempts to characterize human psychological based on what life was like in the Savanna (or whatever environment humans are supposed to be designed by Darwinian forces for) need serious qualification, at best. Speaking metaphorically, evolution is an accident; where "successful", a fortuitous coincidence.In some cases an organism ends up with a set of traits that work out for it in a given environment and it lives long enough to reproduce (even if living in a great deal of pain). Obviously the given environment will impose certain limits, and these limits may lead to certain "modifications" if not extinction. But the assumption the organism is well designed (well adapted, if one prefers secular terminology) for environment X therefore automatically poorly (or less well) designed for environment Y, where X precedes Y, is misleading. Logically, we may be better adapted for our present environment than any previous one we've inhabited (and one can come up with imaginary environments that are far superior than any we've experienced)--it's question we can only answer by looking at X and Y and the organism's traits very carefully, and then perhaps only with a great deal of uncertainty. No one talks about the hand being well adapted to the Savanna and ill adapted the modern city yet analogous arguments re: psychology crop up constantly--esp. re: politics and economics, where extreme irrational prejudices operate.
One more case to consider-- if a country is invaded, it may be less badly off after successful resistance than if it surrendered.
How about WWII? At the end of WWII, the USA was certainly better off. The economic boom we experienced following the war was quite large, not to mention the baby boom.
I still think that counts as lose-lose, though not for the reason daedulus2u gives in the sibling comment. WWII destroyed the productive capacity of several nations and diverted huge amounts of resources to swords instead of plowshares. Had there been no war, these resources could have been committed to making more plowshares, including for the people of the US. Don't mistake a recovery in plowshare production capability, for a higher absolute capability.
Ah, okay. I must admit that the depth of my knowledge on the economic history of my country is rather skin-deep. I do know, however, that we squander unbelievable amounts of money on war, so I think you may easily have a point.
I think people have an amazing ability to leave the costs of war out of their bookkeeping.
That's what the U.S. government does.
I think the 416,000 US military dead and their families would disagree that the war made them better off.
That's irrelevant. Of course you can always cherry-pick people whom some event made worse off. The question was whether the war made the country better as a whole, not whether any individuals suffered.
Actually, what you consider and what you don't in this sort of calculation is an interesting question. Was a particular country better off? Did it work out differently in different regions? How about the whole world? Getting back to the US, was there a cost to the belief that war is good for the economy? Was there a cost to smugness from winning the Civil War and being on the winning side in WWI and WWII?
You're getting into much deeper water here. "What does it mean that some scenario is good/better/best?" is the ultimate, fundamental value judgment. Giving a thorough answer to that question goes a long way towards explaining/understanding yourself, and it's an exercise everybody should do as soon and as often as possible, even though it is by no means easy or quick.
This sounds as though you've worked with that question yourself. What have you learned from it?
There's no way I'm putting it down in a comment, unfortunately - if I do go through the effort of writing down my moral system in a linear form that is understandable to other people, it'll be a several-pages-long essay (possibly a LW post, though). Step zero, for what it's worth, starts with asking "why do I want X?", and recursing that question until you hit an answer you can neither question (without questioning reality itself) nor alter.
Of course I agree with you. I am merely thinking in dollars and cents here, since that is the primary measure of value in the "civilized" world.

Envy is pain at the good fortune of others. -Aristotle

What is the difference between what Aristotle called the vice of envy and the zero-sum bias? I wonder if these two concepts are aiming at the same thing, and if we can learn about one by reading what has been said about the other.

There are certainly many cases where Aristotle's definition of envy would adequately describe instances of zero-sum biases. In particular, when one observes the suffering of others in contrast to those with immense wealth. This can often be seen in comparisons to the west and the third world, as well as in the concepts of "core" and "periphery" in world systems theory. The problem with using the word envy, as I see it, is that the word in its currently accepted form leads people to assume that dissatisfaction is on behalf of the agent alone and not on the behalf of others. It would only reduce the clarity of language surrounding the subject.
I have a feeling most Less Wrong readers dislike using the word envy to describe this, though once can get a acurate description of zero sum thinking by using envy in many context. Your post got me thinking perhaps one could tap into the strong Western arceotpye of "virtues" and things like "seven deadly sins", using them as a pedagogical and memetical tool? I know religious connotations and language are unpopular, but if we find Elizers stories about what are basically romanticized monks (but Bayesian! or in the future! to use a actual TV troope) for ilustrating ideas perhaps compling a list of seven cardinal sins of a aspiring rationalist and matching them with the traditional ones could be something people would be willing to consider. * Gula (gluttony) * Fornicatio (fornication, lust) * Avaritia (avarice/greed) * Tristitia (sorrow/despair) * Ira (wrath) * Acedia (acedia) * Vanagloria (vainglory) * Superbia (hubris, pride) I can see a few posibilities on that list. I think there have been things like calling things rationalist virtues ect. I'm proposing a more explicit piggyback on the existing mems.

Important and interesting theory. Have you read this article by Paul Graham Mind the Gap? It touches on the aspects of zero-sum economic environments, and argues that the big change happened with the creation of the middle class in Europe. This was the first time it was possible to create wealth and keep the profit to yourself. The possibility to accumulate a fortune by creating wealth is just a couple hundred years old.

Maybe it’s a bit on the side but the “Daddy model” in the article seems to me like zero-sum bias.

Thanks for the reference.

See "folk economics" and Bryan Caplan's research for evidence of zero-sum bias. And Robert Wright's Non-Zero for a discussion of how human progress has come from increasingly positive-sum interactions.

This post feels a bit lacking as it just briefly introduces a bias and explains why it is there, but doesn't talk about how to overcome it or the harm that it does. Zero-sum bias results in tons of bad laws, for example. And it can be self-perpetuating - very high marginal tax rates are often justified by assuming the earner just got lucky or too... (read more)

High marginal tax rates are bad only if you assume that the highest earners are most productive. But are they? Don't forget that wealth is also self-perpetuating. If I earned $1 million - whether by hard work, or by luck - then I could live comfortably off the interest for the rest of my life. In terms of satisfying my basic needs, I would have zero incentive to continue being productive. If I live off the interest, am I somehow being productive by proxy of my investment? That's another question.

The ridiculous amount of activity by the untrained and uninformed on the stock markets seems to suggest that people could use a significantly higher amount of "zero sum" bias there. There are tons of people who would not peg their investment skill in the top 1%, who nonetheless think they have a good chance of making money in short term trades. There certainly seems to be the thinking that the stock market is a source of profits for one and all, regardless of investment skill and understanding.

This bias however is most prevalent in the idea that getting rich cannot possibly be a noble act, a point extensively addressed in Paul Graham's essays.

Actually, in regards to your first point, I suspect that people do have an intuitive sense that playing the stock markets is a zero-sum game. I think that the main reason that people exhibit irrational stock market behavior is overconfidence bias.
Great points. I will link this comment in my next post.

Upvoted, as this is an extremely important topic.

I am not sure that zero-sum or positive-sum games translate very well to the hypothetical early human society. You need more analysis to be convincing. What does it mean to have more resources when food cannot be stored and any surplus must be shared with others in the hope that they will do the same in the future? What does ownership or control of resources mean where only what can be carried can be protected? How does game theory cope with gift giving rather than trade? How does game theory work within families? etc.

In a tribe of a fixed small size, food allocation would be mostly zero-sum. Early humans could store meat by smoking or burying it. Also, food surplus wasn't the normal state of affairs: mostly early humans would have been struggling. But sure, there would also be cases where reciprocal altruism worked well, which is why humans have such a keen urge for signalling high moral values and loyalty. I think that what is important is the relative importance of positive sum versus zero sum games in the ancestral environment versus today.

But how do you resolve a situation where there is actually only enough of a resource for one individual or group in a positive-sum manner? The historical record is clear that there are zero-sum situations, and in those situations there can be only one winner.

I don't believe you when you say we live in a world of abundant resources or that we can simply create new resources or that we can design a perpetual motion society of "lasting prosperity".

An argument between you and me over your thesis would not end in two winners. I think it's a silly, ideological, unscientific proposal that is unquantifiable in all its essentials.

An argument between you and me over your thesis would not end in two winners.

It should. When two people disagree, something is amiss, and when the truth is more clearly discovered, both learn something, maybe a little bit for the person who was right, much more for the person who was wrong. And the only way to discover that you are the person who was wrong, the person who would benefit the most, is to make yourself perceptive to good argument.

This is incredibly well-stated; could we get this or something like it added to the welcome post?

An argument between you and me over your thesis would not end in two winners.

If you don't think that arguing about ideas can/should be positive-sum, this may not be the site for you.

The claim is not that all situations are positive-sum, but that people tend to see positive-sum situations as zero-sum because that feels natural and intuitive. You're arguing against a straw man.
and I have a match... as long as situations have the potential to be regarded by the unit of winner (individual, family, group?) as zero-sum then you're stuffed. as I said below resolving a situation as positive-sum just shifts the zero-sum situation to another level: e.g. you vote in a Hawk|Dove leader as President, it's a win-win for sub-state groups but your state loses versus both the Doves and Hawks.
Another thing - what is your unit of winner? The individual? The family? The group? You are assuming scenarios such as Person A versus Person B, etc. What about Person A1 versus Group A which consists of Person Ans when there is simultaneous a game played between Group A and Group B? Person A1 wants to be leader of Group A, and is in a run off with Person A2 for the role. Person A1 is running on making Group A's economy sound but Person A2 promises to protect Group A against threats made by Group B. How can this situation, which is a very common trade-off in politics, be resolved in a positive-sum way? Seems to me you can only work a "positive-sum" solution by sweeping the zero-sum problem to another level of analysis. If the skills of A1 and A2 are used together, that is a positive-sum alternative, but Group A will then not have sufficient protection and lose against Group B - so Group A loses the zero-sum game. If only A1 or A2 become the leader then either A1 or A2, and the alternative trajectories they stood for, will have lost. What about Group A and Group B getting together to decide to share power? Group A and Group B are comprised of Person Ans and Person Bns. Ans and Bns have different life priorities...
You're missing the point. The claim being made is not that zero sum situations don't exist. The argument being made by the essay is that non-zero sum situations exist and that it is a problem when people erroneously label non-zero sum situations as zero sum situations.
All parties who gain or lose in the transaction.
Sorry, I don't understand. Since when was the loser the winner? What I mean is when Individual A and Individual B agree on a solution to ruling State X those INDIVIDUALS are the winners - the individual is the unit of winner. But simultaneously there is a game going on between State X and State Y. The resolution of the leadership contest between A and B may have been positive-sum at the level of the individuals involved but what about at the level of the State? And if it is bad for the state then it is bad for the individuals so what appeared to be a positive-sum outcome was actually an illusion, due to the multi-level problem. This has yet to be addressed by anyone. To recap my very general position: the essay is ideological, not scientific. The title says everything.
What does this mean? What do you mean by ideological and what do you mean by scientific? The point of Less Wrong is to be, well, less wrong. Is that an ideological goal? We like reducing cognitive biases and getting a better understanding of reality. Do you consider that to be ideological? And if this essay is ideological rather than scientific, why does that matter?
i misunderstood your post. The unit of winner is an individual - that seems pretty obvious in this context. Organizations don't act, only individuals. What I was talking about (and what I thought you were asking) was the group of people you use to judge whether the game is zero sum or not, and that is... All parties who gain or lose in the transaction. Just no. If hurting the state helps one individual gain more power over the state, it can certainly benefit that individual.
Can you not see the irony in the title of your post: "Fight Zero-Sum Bias"?

The only person you should try to be better than, is the person you were yesterday

Our ancestors hunted together and lived in groups. If it had been a zero sum game there would be no point in living in groups.

See a comment by Roko which addresses this point and which I agree with.

The zero-sum bias seems to be also responsible for the the concept of karma which is a ubiquitous concept (not just amongst Hindus). The roots of this can be found in the ancient religious texts like the Bhagvad Gita and go on to support what multifoliatrose says in the post.

But if karma was how the world worked games would tend towards having higher, not lower, absolute sums.

Not to be a cynic, but:

Zero-sum is not an illusion.

Every single thing has opportunity cost.

It may not be another person that pays if I pick up my toys, and go found my own community elsewhere.

It may be fewer trees and less open land.

We're all in this together, on the same planet, sharing the same air and water. Nothing is positive-sum if it involves using physical resources.

I have water, but no air. You have air, but no water. I give you water in exchange for air. Presto! Everybody wins.

With no mention of the evolution of empathy, compassion and altruism, (let alone no mention of George Price's work or Frans de Waal's insights), I'm unsure that you can draw the conclusion we move upwards from a zero-sum bias.

In fact it is our adaptability across a range of elements that gives us the ability to be zero-sum, -two-some, and more-some.