The question was generated by ChatGPT. I took the liberty to simplify it just a bit. More details below.
This market will resolve as follows:
If on Feb 1st the market is trading in the range of 35-49%, the market will resolve to NO
If the market is above 49%, it will resolve to NO as well
If the market is below 35%, it will resolve to YES
Now figure out a strategy to make profit here!
P.S.
Just to make the idea more transparent.
35-49% is technically below 50% so it will resolve to NO for correct people to profit.
Below 35% will resolve to YES so people who downvoted too much end up losing Mana. And the other way around, if it’s above 50%, resolving to NO will make people who upvoted too much lose.
1,000
3.00@Soren But if we take 50% as a baseline, going for 42% would require you to vote NO, and I assume you’d like to win Mana from your bets, not lose it. If you voted Yes, this would take the market away from the target range, thus making it less likely to resolve in your favor.
@Soren And paradoxically voting Yes is feasible when the market is below the target range (like it is now), because if it stays below the range then the Yes team will win everything.
@Soren But of course by voting Yes you would also be taking the market into the NO territory potentially ;)
@JosephS “But if we take 50% as a baseline“
The word “if” here is doing a lot of heavy lifting. Why would we take 50% “as a baseline?” What does that even mean? As it stands, the title and question description contradict themselves.
@mathvc Thank god i didn’t participate in this market before looking through comments! What a waste of human time.
@mathvc Bro, logic and humor are very wasteful, unlike betting on “Will Trump win the US election”. Luckily you didn’t fall for it, what a relief. Hope admins ban me from posting anything challenging again.
@MingweiSamuel The idea is that 35-49 is technically below 50% so it should resolve to NO for correct people to profit
Below 35% will resolve to yes so people who downvoted too much lose Mana. And the other way around, above 50% will resolve to no so people who upvoted too much also lose.
@JosephS This makes absolutely no sense. The question is "was GPT correct" which it would be if the market was between 35 and 49. It doesn't matter if that's below 50%. It wouldn't matter if the market needed to be at 0%, it would be YES because the answer to the question is YES.
@Pykess You just assumed the market would resolve however you interpreted the question without even reading what the author wrote about how it would resolve?
It’s a fun self-referential market, the author can make the rules whatever they want, it doesn’t have to make sense
@DylanSlagh - I think the description of the market and the title of the market, should broadly agree.
The title is written as a yes/no question, but the market doesn't resolve to the answer of that question.
There is nothing inherently wrong with the self-refferential and abstract resolution criteria in the description, but it doesn't align with the title of the market.
For instance, if I somehow knew for a fact that indeed it is true that "ChatGPT will have correctly predicted the outcome of this market at 42% (with 7% error margin)", then instead of buying yes/true shares, I should buy no/false shares.
It is like having a market that says "Will Joe Biden win the election?" and it resolves to "no" if Joe Bien wins. I don't think being self-refferential is relevant to the complaint here.
@JosephS -
>"The idea is that 35-49 is technically below 50% so it should resolve to NO for correct people to profit"
Can you try to explain that again? Why is that relevant? How does this point to the correct people?
>"Below 35% will resolve to yes so people who downvoted too much lose Mana."
But they were correct to downvote? In the below35% scenario, ChatGPT failed to predict the market (within 7% margin), so "No" is the answer to the question, and the downvoters .
>"And the other way around, above 50% will resolve to no so people who upvoted too much also lose."
Makes sense.
-
The "below 50%" thing doesn't make sense.
Resolving "No" doesn't cause people who voted it to within 35-39 to profit.
For instance, if the market was at 34%, and the last person to bet bougt Yes shares to bring it to 35, your market takes all their money and gives it to others.
@DylanSlagh For one, the title and description should agree, regardless of the market. And two, using your logic all of Levi Finklestein's questions should've been a-okay and he should be unbanned.
@MatthewLeong @Pykess Regarding the title, you’re welcome to suggest a better one, however consider there’s a length limitation. Please note though that the title says “Will have”, not “will”, so technically it’s correct. The actual proposition here is whether the humans will collectively align themselves with an already made AI “decision”, not whether that decision was correct.
@MatthewLeong @Pykess Now regarding the reward scheme. I’ll just say perhaps you didn’t fully think through the idea. I don’t want to explain my logic yet because people should figure out their own trading strategy. The bottom line is that in a capitalist market the motivation for trading should be to profit, not to answer “yes” to a sentence that naturally invites that answer.
A hint for you: remember that anyone can bet multiple times.
Also note after 50 trades already made by now the outcome stands exactly at 42%. Whether it’s a testament to people’s motivation to profit or to the meme nature of the question is just another factor for you to consider when placing bets.
@Pykess @MatthewLeong I updated the title to say “Trick Question” instead of “Humans vs AI”. Maybe this will result in less people blindly voting based on the title.
@JosephS if lots of people bet lots of mana, isn't there going to be an easy profit possible from last minute no betting a large sum
@JosephS >The actual proposition here is whether the humans will collectively align themselves with an already made AI “decision”, not whether that decision was correct.
Sure, but you can buy shares that align with that decision, and lose mana.
For isntance, at time of writing ithe probabiltiy is 36%.
Suppose that I buy "YES" to bid it up to 42%, as predicted by ChatGPT. I will have helped make ChatGPT (more) correct. If no one else bids after that, then you'll resolve the market to "NO" to take all my mana.
Now, I can read the description and see that you have not made the question such that you want me to big to make the AI more correct.
But don't then say that you've set up the resolution rules to align with the AI's decision.
You might think that on aggregate, the people that bought no shares helped it get to under 50, and hence closer to the 35-39 range.
However depending on the price at which people bought their shares, that might not quite be true, and aggregating them doesn't make
Basically, your prompt has nothing to do with what ChatGPT has said.
ChatGPT spat out a nubmer, 42%, and then you made an quetion with a title about ChatGPTs prediction, and then made resolution criteria that bear no relation to the prediction (actually is contrary to the predicition).
If you wanted to amke the 'correct' people profit, you could have made it resolve to 47%. If I understand the mathmeatics correctly, then everyone who bid it towards 47% would profit (although telling us that you'd do that would make the market easy to bid on).
@MatthewLeong Actually I agree with most of what you said. The most basic version of this would be one of those markets that resolve to no above 50 and resolve to yes below 50. What I was trying to do is to make that idea more interesting by inserting a range in the middle where people could make a profit. I agree that not everyone betting in the right direction will make profit, but on average I believe they will.
An additional interesting factor here is that if everyone decided to bet NO, then the market would go below 35% and that would make the YES team win big time. So it’s actually possible for the yes team to win even though the obvious conclusion most people should make is that NO is the only viable bet here.
The end of your comment where you talk about 47% doesn’t quite align with my own math, but I agree this market is practically destined to resolve at a particular specific number. In some cases I might technically be able to make the last bet to get that number, but that depends on how much Mana is invested vs. my budget.