TLDR simplification:
YES = prediction markets good, pro-social on balance.
NO = prediction markets bad, anti-social on balance.
(Sorry for the confusing double-negative wording of the question)
It seems true to me the observation that many others have point out that there is an inherent connection between prediction markets and motivation. If someone has a stake in a market about future outcomes, they have motivation to affect those outcomes if they can.
If someone sees a market they believe they could affect, then it is a temptation they are faced with to enter into the market and deliberately affect the resolution.
Sometimes people describe this phenomenon as 'all prediction markets are assasination markets' to remind users that this phenomenon can be manipulated to bring about harmful antisocial outcomes in the world. I prefer the more neutral stance, since it is an effect which can equally easily be used for good or ill. It is not an tool biased towards negative uses, like a dirty bomb specifically designed to harm civilians. It is a neutral tool, a force for good or ill. Let us use it wisely.
Resolution criteria: If, during the open period of this market, I am presented with sufficiently persuasive evidence that the defence/offence balance of the motivational power of prediction markets is weighted towards anti-social ends then the market will resolve NO after a discussion period of approximately 6 months. [Edit: I made some confusing statements here about waiting for a six month countdown because I was worried the question would get no engagement. The question got plenty of good engagement, so the six month countdown has begun.]
The market resolves YES if it reaches the closing period without me having been persuaded that the persuasive power is more potent for antisocial uses.
Related questions
🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ267 | |
| 2 | Ṁ38 | |
| 3 | Ṁ37 | |
| 4 | Ṁ32 | |
| 5 | Ṁ25 |
On PredictIt, when they had the markets on the 538 average presidential approval at the end of each week, we were able to prove that the market flipped on the last day at a higher rate than should have been able to happen by chance.
It turns out that this was because a number of people who had been playing that for a long time would seek out obscure polls and report them to 538 as late as possible, in order to affect the outcome of the market, but particularly in such a way to allow them to change the outcome to a bracket that intrinsically had a low probability (i.e. that would not have happened at all, apart from their actions.)
I reported this to 538 and they said, more or less, "yep we agree that is bad behavior but there is nothing we can do about it," which of course was a falsehood; they could have delayed the inclusion of those polls by 24 hours, which would have happened anyway in the absence of attempted manipulation.
Depends upon how the market is laid out and run.
To that end, I have put together a grant to hopefully encourage prosocial markets.
https://manifold.markets/PatrickDelaney/interesting-markets-grants-submit-h
Lsusr has written a post or two about this question: https://www.lesswrong.com/posts/D5Dq9AyXqYfhLiaC3/prediction-markets-are-for-outcomes-beyond-our-control
@NicholasKross Thanks. Quoting tailcalled's highly relevant comment on the linked post here: "tailcalled 1y 13
If it gets to 99.9%, you STILL can't make a profit by switching sides - you're too heavily invested on the "yes".
Why not? If you sell your yes shares, that seems like a guaranteed profit to me? Are you assuming very large transaction costs or something?
Like let's make it dead-simple:
The market sells contracts that pay $100 if you make the post, and pay nothing otherwise.
Initially the market price is at $1. This means you can earn $99 by buying a contract and the making a post, which sounds like a cost-effective way of spending your time, so you buy a contract. This increases the price, maybe to $2, maybe to $50, depending on the liquidity. (It sounds like you prefer the high-liquidity limit?)
The market somehow finds out about your plan, maybe by watching the prices, maybe because you announced your intention on social media, maybe for some other reason, so they conclude that you will make the post and therefore bid up the contract to $99.
Now you could make the post and earn a total profit of $99. But you could also just sell the contract at the current market price, yielding a guaranteed profit of $98, which is less risky and requires less work.
It sounds you are saying that this story fails at step 4, because you are saying that one couldn't make a profit in this circumstance due to being "too heavily invested on the "yes"". But I don't see how you're too heavily invested in the yes; it's true that you've got a "yes" contract, but you bought it at $1 and the market price is now $99, so you can sell it and make a profit of $99-$1=$98.
"
Confusing double-negative wording in the question title: https://languagetool.org/insights/post/double-negatives/
Some great comments on here so far. I'm delighted by the thoughfulness and thoroughness of your answers. For clarification here, it seems to me that as of 2023-03-18 the debate is leaning heavily towards YES ( the world-optimization force of prediction markets is neutral or more helpful for good things) and away from NO ( the world-optimization force of prediction markets will tend to favor destructive ends because of offense-defense imbalance as mentioned by Steven.)
Summary so far
Some points in favor of YES (prediction markets neutral or good overall)
Interacting with the prediction market is by nature a public act, and thus unlawful actions which benefit a particular buyer's position will cast suspicion on that buyer. Risk of law enforcement may thus restrain bad actions.
Some points in favor of NO (prediction markets bad overall)
If lots of people are using it, and many of them are kinda selfish and willing to take low-effort low-risk actions to gain money if those actions aren't obviously very bad, then there will be a preponderance of mildly negative things (below the notice of Law Enforcement or not even illegal). If there's a lot of such actions, chosen for surprise-thus-profit not necessarily negative outcome, and the offense-defense imbalance will cause the overall outcome to be negative. A neutral optimization force for enacting sneaky surprises pushed through a filter.
A question in my mind:
My view is that the majority of humans act in mostly pro-social ways most of the time, especially when there is an orderly society with a rule-of-law in place making outwardly anti-social behavior punishable and thus less rewarding in expectation.
People clever enough to come up with sneaky novel surprising ways around the law enforcement / social-pressure enforcement / childhood moral conditioning often choose to only deviate in relatively minorly-harmful ways (embezzeling money rather than shorting stocks on a large company before lacing their offices with anthrax.)
Most people like to be able to tell stories about themselves where they are the heros, not villians. Will the optimization pressure towards being able to surprise the world be a sort of end-run around human moral intuitions via the offense-defense balance filter? Will it get people to do more anti-social things on average than they would otherwise have done?
Note: I meant 'assasination markets' as an extreme example, a stereotyped stand-in for 'anti-social effects', not to mean that that was the only bad thing the question was considering.
Let's say you need to achieve X motivation points to do something (marginally, given other existing factors). You can see it as how "easy" it is in a sense which seems more relevant than the usual meaning of the word.
A prediction market gives you ε points (a small amount). People are already quite motivated to do good things, so ε is enough in a significant minority of such cases. This is almost never true for bad things, and you have to have some unusually strong reason to do them that happens to be just not enough on its own.
I think, in general, this force is neither good nor bad; it's neutral. It encourages people to do things that diverge from what most people believe, but that doesn't make them good or bad.
Assume, for the sake of argument, a society with widespread use of prediction markets, including real-money markets, that most people pay attention to. Thus, assume that those markets reasonably accurately reflect reality, and that they're an effective communication mechanism for what many people believe, and that there may be enough money to be made to serve as motivation.
In general, given those assumptions, a prediction market should do a good job of making information available. If a market (given the above assumptions) is turning out wildly different than you'd expect, you should be asking "what do people know that I don't". It can only add information people wouldn't otherwise have had.
Any time you know something people don't know, you stand to make money in a prediction market; in an ideal prediction market, that's the only way to reliably make money. Given that, the phenomenon we're talking about here is that if you think you can affect a market, in ways that people wouldn't expect you to, then you stand to know something other people don't: your own high-impact future behavior, or willingness to engage in such behavior in the future.
Even in the context of markets commonly expected to produce negative purposes, such as "will this person die", they're communicating information that can be used for positive or negative purposes. If you discover that wildly more people expect you to be dead than you would have thought, then 1) you should find out what people know that you don't, rapidly, and 2) you should consider some personal security measures. You might not have done either of those things in the absence of the market. The net effect was the conveyance of information that wouldn't otherwise have been conveyed.
Conversely, if you have a fatal disease and there's a "will you die by X" market, and based on priors for survival rate of that disease people predict that you will die, that could turn into a "cure the disease" market.
If there's a market about whether a company will succeed, and it's showing that most people think the company will succeed, you stand to profit if you can cause the company to fail. If there's a market about whether a company will succeed, and it's showing that most people think the company will fail, you stand to profit if you can cause the company to succeed. And whether any of those actions is good, bad, or indifferent, depends on how you feel about the company.
In general, I do think this phenomenon fundamentally represents the ability to diverge from what most people expect or believe. In that specific regard, I think this is "not social" in the sense that it represents a person or small group going against what most people believe. However, I don't think it motivates strongly towards "good" or "bad" actions, just actions that diverge from what most people expect. Those diverging actions could be bad or good.
So, the question becomes, do you believe that the ability for individual people or small groups to diverge from widespread expectations is on balance good or bad or a balance of both?
I think the ability to do better actions than most people expect and be rewarded for it is a major driving force for startups, innovation, creativity, and similar. I also think the ability to do worse actions than most people expect is a source of major conflict. That alone doesn't argue for one of these winning out over the other. So I'll end with one thing I think serves as a balancing force:
It's easier to predict successful negative actions that require little innovation, than it is to predict successful positive actions. If an action is easier to predict, there will be less money to be made in diverging from expectations.
@blake That displays a strikingly naive view of the effectiveness of police - it seems like you're implying they're nearly 100% effective at solving these kinds of crimes. State-level actors, at least, are often able to assassinate people (Bin Laden, the Iranian general, Litvinenko. etc.) with a low risk of prosecution/retaliation. Even criminal gangs or individuals might be able to get away with it - not that I'm recommending this.
Also, not everywhere has police. At one point in time, in the recent past, Somalia didn't even have a government, let alone police.
@RobinGreen I assume if it is a death market big enough to get frequent attention by manifold markets, it probably also has good enough security
There is already a market for most antisocial behaviour call real life.
There are much easier ways to profit from antisocial behaviour than prediction markets (which is why we have to socially discourage the behaviour in the first place.)
On the other hand, we don’t have great ways to encourage pro social behaviour for good ends rather than just virtue signalling. Look at the most popular charities and how much they spend on marketing and making donors feel and look like good people.
To put it another way: the impact at the margin of prediction markets on antisocial behaviour is minimal. If you didn’t already have a reason to assassinate someone, a prediction market where you are very clearly indicating your interest is not going to push you into actually doing it, as there are still lots of reasons not to.
Conversely, the ability to gain reputation and even potentially profit from prosocial behaviour is augmented by prediction markets. It may be a small effect but the antisocial effect is so close to nonexistent it still wins.
@Nps It sounds like you should be predicting Yes then. Did you misread the market description?
@RobinGreen I just reread the title and description and still can’t parse which way Yes and No are meant to go. Too many negations
@Nps you can treat it like an algebra problem and remove the double negation, turning it into the easier to understand "All prediction markets are motivation markets. Is it true that this force is not proportionately more useful for social ends (assassination markets) than pro-social ends (charity impact markets)?".
Is Will Putin be the President of Russia at the end of 2023 an assassination market, or a market on whether the Russian people will riseup and overthrow him with nonviolence as we've seen happen countless times (notably the fall of most of the former Warsaw Pact "Communist" governments?
Even if it was an assassination market, is it evil? (Probably, because whoever takes over is likely to be worse - but I'm guessing a lot of Ukrainians would be willing to risk it).
Prediction markets are weighted towards evil because most people in the real world are trying to do good, so evil is almost an underexploited economy when markets arise that allow for it. Plenty of people are trying to solve alignment (so prediction markets are a pretty small factor in that task), not that many people truly feel an incentive to do something horrific
@LukeBousfield I think the answer may depend on the scale. A bunch of smaller markets like Manifold are weighted towards good, because most people are trying to do good with it and there's just not much to be gained in exploiting the system. If you've got a large sample of people working very large markets, however, you will DEFINITELY get a few folks trying their damnedest to bend its power to get what they want at everyone else's expense.
something something decentralization, clap clap clap
I feel like compounding and signaling motivation towards positively received goals is more beneficial to one's status in most contexts.
I therefore believe that the motivational impact of prediction markets on "positive" goals is stronger within the narration contexts of human status games.
There is some similarity to the argument/belief about Adam Smith's idea of the markets' invisible hand to drive wellbeing of societies here.
As we are dealing with meta markets of predictions and not 1st order goods/financial markets, I would also guess manipulative power dynamics by extreme sociopathic egocentric behavior to be more modest for predictions.
Consider the greater context. It's both illegal for a private individual to put a bounty on someone else's life, and illegal to kill someone. The benefit of a11n markets is acausal anonymous coordination, but it's as hard to kill someone and not get caught as before. Also, there are already plenty of anonymous ways to offer darknet bounties. I'm not sure that prediction markets adds additional incentive.
@noumena I think you're right about assassination markets and other very illegal things, but there are lots of negative value actions that aren't illegal, or are illegal but are easy to get away with.
This market allows the possibility that the market maker didn't intend to incentivise action but did do anyway. That's what I think is a risk.
A simple example might be that somebody uses Wikipedia as a resolution criterion so somebody else edits Wikipedia to win the bet. That's pretty minor, but it's still antisocial.
@noumena It's apparently legal to go to Ukraine and kill Russians. And also to crowd source funding for this (in the US).
@AaronKreider Yes but killing Russian soldiers who are engaged in invading Ukraine isn't "antisocial". Nor would it usually be described as "assassination", unless it was a particularly high-value target I guess.