4 ms·

This was also featured in Scott Alexander's Substack [1] earlier this week on Monday. The really bizarre part about this platform is that not only is the person

by usrme 5y ago

This was also featured in Scott Alexander's Substack [1] earlier this week on Monday. The really bizarre part about this platform is that not only is the person posing the question responsible for judging the outcome, they also get a percentage of the trading volume as a fee [2]. To me, this makes it way too susceptible to influence and be too subjective to even take seriously when compared against platforms like Metaculus [3]. --- [1]: https://astralcodexten.substack.com/p/mantic-monday-dogs-in-wizard-hats https://astralcodexten.substack.com/p/mantic-monday-dogs-in-... (see "Mantic Everyday" section) [2]: https://mantic.markets/about https://mantic.markets/about (see "How are markets resolved?" section) [3]: https://www.metaculus.com/ https://www.metaculus.com/

nabla9 5y ago
Generally these open to everyone markets suffer from the garbage in garbage out problem (Good Judgment® Open, Metaculus,..) The low quality of the questions indicates that nobody spends any effort. The question may be receiving community input for a week and nobody has a clue. Metaculus admins especially are spamming (their own site) with confused low-effort questions. They have: 1. arbitrary ways to quantify outcome thresholds. No standardization across questions. 2. Writing questions using terms the author does not actually know or misunderstands. 3. Using some random news sources as resolution criteria. Even if they are generally reliable, it just reveals how the author really didn't get into the question. For example, war or conflict questions. A question may use some arbitrary number like 3 or 5 dead as a threshold for war. Maybe trying to figure out how subject matter experts are quantifying different conflict thresholds and what their reasoning was would be reasonable.
- namenotrequired 5y ago
  Wait, are you saying metaculus admins are spamming their own site or their tiny competitor’s site?
  - nabla9 5y ago
    No. They are spamming their own site with low quality questions.
PragmaticPulp 5y ago
> The really bizarre part about this platform is that not only is the person posing the question responsible for judging the outcome So the person posing a question can arbitrarily choose the answer that benefits themself financially? Just click the button that says the outcome you bet on was the correct one, then collect the money? This seems like a non-starter for any platform that wants to be taken seriously. What am I missing?
- skrebbel 5y ago
  I think what you're missing is the reputation of the person who poses the question. I think the idea is that you don't get questions like on Quora, asked by internet randos. Nobody would bet on those, for exactly the reasons you brought up. But someone who already has a following, someone who has a strong reputation in the field that the question is about, would create quite the controversy and lose much of their audience if they'd first strongly signal boost their question among said audience, and subsequently pick the wrong answer. So I think the idea is, when choosing what questions to bet on, you're not only thinking about the question itself but also about the trustworthiness of who poses it. A bit like Kickstarter in a way. That makes it part of the risk, maybe even part of the fun. This also means that this service will fail if they don't manage to attract sufficiently high-profile people to pose questions and signal boost them.

by usrme 5y ago

nabla9 5y ago

Generally these open to everyone markets suffer from the garbage in garbage out problem (Good Judgment® Open, Metaculus,..) The low quality of the questions indicates that nobody spends any effort. The question may be receiving community input for a week and nobody has a clue. Metaculus admins especially are spamming (their own site) with confused low-effort questions. They have: 1. arbitrary ways to quantify outcome thresholds. No standardization across questions. 2. Writing questions using terms the author does not actually know or misunderstands. 3. Using some random news sources as resolution criteria. Even if they are generally reliable, it just reveals how the author really didn't get into the question. For example, war or conflict questions. A question may use some arbitrary number like 3 or 5 dead as a threshold for war. Maybe trying to figure out how subject matter experts are quantifying different conflict thresholds and what their reasoning was would be reasonable.

namenotrequired 5y ago
Wait, are you saying metaculus admins are spamming their own site or their tiny competitor’s site?
- nabla9 5y ago
  No. They are spamming their own site with low quality questions.

PragmaticPulp 5y ago

> The really bizarre part about this platform is that not only is the person posing the question responsible for judging the outcome So the person posing a question can arbitrarily choose the answer that benefits themself financially? Just click the button that says the outcome you bet on was the correct one, then collect the money? This seems like a non-starter for any platform that wants to be taken seriously. What am I missing?

skrebbel 5y ago
I think what you're missing is the reputation of the person who poses the question. I think the idea is that you don't get questions like on Quora, asked by internet randos. Nobody would bet on those, for exactly the reasons you brought up. But someone who already has a following, someone who has a strong reputation in the field that the question is about, would create quite the controversy and lose much of their audience if they'd first strongly signal boost their question among said audience, and subsequently pick the wrong answer. So I think the idea is, when choosing what questions to bet on, you're not only thinking about the question itself but also about the trustworthiness of who poses it. A bit like Kickstarter in a way. That makes it part of the risk, maybe even part of the fun. This also means that this service will fail if they don't manage to attract sufficiently high-profile people to pose questions and signal boost them.