Prediction markets and their role in predicting the election

January 8, 2025 7:35 AM

Prediction markets can make a valuable contribution to policymaking. Besides elections, prediction markets are also used to bet on other activities like sports outcomes. Social prediction markets like Manifold allow users to create their own markets for others to bet on (albeit with play money)
posted by buffy12 (20 comments total)

On the Manifold front page “did Covid-19 come from a laboratory” - 47% yes - close tab.

This is some bullshit right here.
posted by seanmpuckett at 7:44 AM on January 8 [7 favorites]

Mod note: Added 'uspolitics' tag so that users can choose to see this post or not via MyMefi
posted by Brandon Blatcher (staff) at 7:50 AM on January 8 [1 favorite]

Finally, a key advantage of prediction markets – which may in fact partially explain why they have not caught on – is that they upend hierarchies. They challenge participants to put their money where their mouths are, regardless of rank or reputation.

They also offer new opportunities to manipulate policy for fun and profit.
posted by RonButNotStupid at 8:28 AM on January 8 [8 favorites]

They don’t upend hierarchies, they force more and more aspects of public life to conform to the overarching hierarchy of who has more money
posted by Jon_Evil at 8:31 AM on January 8 [9 favorites]

Soooo...this capitalism, it can do anything?

(and, the people who wrote this article, can they frame it to include Gamestop's stock price in 2021?)
posted by AzraelBrown at 8:37 AM on January 8 [1 favorite]

Wow, nothing but negative comments on this thread, starting with the first.

buffy12, I thank you for this post. Prediction markets are fascinating. I ran a small one for several years and the results were useful.
posted by doctornemo at 11:19 AM on January 8

I thought a big problem with prediction markets, especially political ones, is that there's a lot of incentive for participants to try to game the results to use as PR for their chosen candidate/policy (.e.g, place massive bets on Trump skewing the predictions so you can turn around and point to the huge prediction advantage as proof of Trump's strength).
posted by star gentle uterus at 11:43 AM on January 8 [5 favorites]

Goodhart's law has come for the prediction markets. In isolation, a market where all the participants are trying to make money based on what they know can produce surprisingly accurate results. But you can't actually do anything with those results, because as soon as people realize what you're doing, you've effectively put your policy decisions up for auction.
posted by echo target at 11:47 AM on January 8 [9 favorites]

Goodhart's law: “When a measure becomes a target, it ceases to be a good measure.”

When you take a metric and start tying it to IRL outcomes, people will start gaming them both deliberately and subconsciously
posted by Skwirl at 12:18 PM on January 8 [1 favorite]

That is one problem, star gentle uterus, and it takes some design and moderation to mitigate.

But you can't actually do anything with those results
Unless you do. Historically companies and governments use prediction markets to get insights for planning. We did the same.
posted by doctornemo at 1:28 PM on January 8

Could you expand on how your system dealt with that problem? It seems inevitable that people will bet, consciously or not, on what they want to happen rather than what’s most likely.
posted by star gentle uterus at 2:30 PM on January 8

On the Manifold front page “did Covid-19 come from a laboratory” - 47% yes - close tab.

This is some bullshit right here.

Meh, I'd say "around 50/50" is pretty much conventional wisdom among scientists (as far as I can tell).
posted by kickingtheground at 2:46 PM on January 8 [2 favorites]

Metafilter: Wow, nothing but negative comments on this thread, starting with the first.
posted by box at 3:20 PM on January 8 [5 favorites]

star gentle uterus, we did a few things.
1: used play money, not redeemable for legal tender.
2: had a diverse user base which couldn't agree on anything, so different takes tended to balance out
posted by doctornemo at 5:42 PM on January 8 [2 favorites]

One fun thing about prediction markets is watching people who are very confident in their wrong ideas lose money repeatedly. It’s like a tax on engineer’s disease.
posted by mbrubeck at 9:58 PM on January 8 [1 favorite]

Meh, I'd say "around 50/50" is pretty much conventional wisdom among scientists (as far as I can tell).

After quite a while spent looking into this question, I found this survey of scientists from last year. The summary is an easy read, but to quote a few key lines:

"The study’s experts overall stated that the COVID-19 pandemic most likely originated via a natural zoonotic event"

"about one fifth of the experts stated that an accident was the more likely origin"

"The experts mostly expressed the view that more research on COVID-19’s origin could be of value. About half of the experts stated that major gaps still remain in the understanding COVID-19’s origin, and most of the other experts also stated that some research is still needed."

Caveat: this conclusion matches my preconceptions so I may not have examined the study as well as I would have if I disagreed with it.
posted by Bryant at 2:08 AM on January 9 [4 favorites]

Meh, I'd say "around 50/50" is pretty much conventional wisdom among scientists (as far as I can tell).

You can always say something is 50/50 and pretend you were being wisely astute. And your half-of-scientists stat is completely wrong. (When it comes to racist right-wing conspiracy theories, I choose bluntness over tact: your claim is not only irresponsible disinformation, but factually not true.)

Wow, nothing but negative comments on this thread, starting with the first.

Because it's not a good article? This is just a retread of mid-2000s "wisdom of crowds" except with a capitalist bent (more money gambled = more weight) instead of democratic (more gamblers = more weight).

In the bigger picture, this reach towards "what people are saying" as a way to predict reality is a continuation of the rolling back of the Enlightenment and journalism: reporting on people's opinions and an appeal to authority instead of trying to determine the actual truth.

I asked ChatGPT to generate an article based on the headline, and it gave a few hundred words of similar pablum: "prediction markets can predict things accurately, except when they don't." This article is basically content farm slop. Give it a pass, if you're reading these comments before the article.
posted by AlSweigart at 7:05 AM on January 9

This is just a retread of mid-2000s "wisdom of crowds" except with a capitalist bent

Not as such. Modern prediction markets go back to the 1980s so their capitalist bent precedes wisdom of crowds discussions. Or perhaps you and the other negative commentators are more concerned about the markets themselves, rather than the article?

this reach towards "what people are saying" as a way to predict reality is a continuation of the rolling back of the Enlightenment and journalism: reporting on people's opinions and an appeal to authority instead of trying to determine the actual truth.

Fascinating idea. Where do you see the appeal to authority, in the voice of a given prediction market or poll?
posted by doctornemo at 2:22 PM on January 9

The authority is the price. If only my reputation was at stake, say a Brier score with it's punishment for wrongness, and there wasn't a reward from the bet, it wouldn't be a prediction market and wouldn't be about finding someone less informed* than you and taking their money, points or tokens. "If you can't see the chump in the deal, it's you" goes the saying and, here, grifters can find arrogant marks like shooting fish in a barrel.

The other comments about "metrics beginning goals" don't sit right -- I don't disagree -- but it doesn't seem to match the scenario unless also you're taking seriously the stories about demand for positions matching fair polling of the whole population. You can make multiple logins, enough capital can endure losses that most families (or their lines of credit) can't, you can create hype in back-channels and induce demand for unrealistic positions, and the people who stake confidently that they understand the world and can predict it well have both pride and money at stake so can be fooled into doubling down on a losing position.

There's a huge contrast between sports gamblers or prediction-market users with open registrations for logins against the vetted and trusted members of national security infrastructure -- the former aren't playing to keep life from being lost to violence against the state and her people.

I don't find this article persuasive.

*: market forces are said to be fair and to enable rational choices when two parties with full or equal information meet to make a deal -- something not seen in nature or civilization.
posted by k3ninho at 2:30 AM on January 10 [1 favorite]

Further thoughts emerged with "the wisdom kf the stairs" -- if your organisation gains from an internal prediction marketplace, you want to look at the resource allocation and rebalance to account for decision-makers and domain specialists not being the winners in this prediction market.

I'd like to know why aren't your specialists driving the business in the direction that they understand to be optimal? Who else in the business are outscoring them and why?
posted by k3ninho at 6:41 AM on January 12