11/4/25 LAGOS

It's the name of a harbor in portugal

11/4/25: SIXVN

Does this market encourage cooperation or not?

Clearly, yes holders want to cooperate to solve things, so it gets solved & they get paid out.
But yes holders also take from the group (we get fewer hints because you're holding yes.)

If there was a way to fairly divide up the gains from solving, that would be nice. Say, a group of users agree to all hold "no" for a long time, and if it gets solved, they all get paid out.

But the problem is, anyone who learns that the group is about to solve it, could swoop in and buy all the yes and then the group wouldn't get any payout.

The group can't capture gains from their actions - hints are shared with everyone. If the group is about to solve it, anyone else might solve it too.

So... The market only really incentivizes happening to be the one who looks at all the comments so far and solves it solo.

I wish we had all agreed to Yudkowsky's proposed decision theory. I'm reading his Asmodeus thing now. In that world, the solo-solver would recognize & quantify the impact from all the no-holders, and sends mana to anyone who helped, after the fact. But I don't think we've agreed to that & I don't think we can expect smart strangers to agree to that. I don't expect the world to actually work the way yudkowsky wishes it did, so I think we're actually just in a world where we don't solve the problem 😢

@DannyqnOht I wonder what happens if we made a mirror market for this that doesn't gate the hint drops.

@Quroe I think everyone just bets "yes" until the 1000m liquidity is gone, and one of them has good reason to solve it.

@DannyqnOht On this market? Or a mirror market?

@Quroe On your proposed mirror market where info is just provided no matter the odds.

Oh, wait, do you mean literally another market "This resolves yes if the underlying market is solved"?

I have no idea.

@DannyqnOht Just to make sure we're both in agreement on what we're talking about...

My proposal is to make another market that simply asks how this market will resolve, and just copies the result over and resolves the same way.

This could give people an outlet to bet YES without stifling the hint drops themselves, but it also just provides an arbitrage target for risk free bets.

Is this what we are both talking about?

@Quroe I now understand that to be what we're talking about; I didn't at first, but now we're on the same page.

I think all that effectively does is add liquidity to this market though. That seems to follow pretty quickly if we just assume reasonably good arbitrage players are floating around (which I'd personally do if no one else did).

I think it still comes back to the issue that there's no way for active puzzle-solving actions (eg betting no, on either market; sharing thoughts, on either market) to cause profit to the actor.

Yudkowsky's multi-agent theory stuff kinda addresses this (we all decide to solve it together and pay out fairly after the fact to anyone who helped, in a well-defined-but-complicated-to-measure-in-reality way) but I don't think an extra market solves it. I'm curious if there's another solution, beyond yudkowsky's solution, but I'm not that hopeful.

@DannyqnOht I have found that collateral staking strategies have been a useful tactic for garnering cooperation. I was an executor to one of these recently, and it worked well!

In essence, parties stake a chunk of mana and promise to cooperate, on pain of losing their stake if they don't.

The question is how we can adapt that strategy to this format.

@Quroe I think contracts can work if you have very few players, or have reason to expect that random external parties won't muck things up. Or if there's a way to capture gains from cooperation.

But in this case, any random stranger can come in and steal all the winnings from any cooperating group. And there's no way for anyone to stop that.

Like, if we could all beat up anyone who bet yes, then we could sign contracts and actually solve the thing and get payouts, with a designated "yes" better for the very end, who then sends mana on to everyone who helped.

But literally any rando can come up and say "yeah they're gonna solve it" and then the price is 99c and everyone who worked to solve it gets nothing.

@DannyqnOht that's easy to punish. rando who says "they're gonna solve it" isnt trying to solve it themselves. you just eat their YES and stop guessing

@DannyqnOht This could devolve into some of the biggest YES holders offering a lottery to NO holders that pays out if the market resolves YES. That might be a way to allow people to hold NO while still having a stake in YES, virtually pushing the gate to hint drops around as the EV scale shifts.

@Stralor Ok, worse scenario: The good guys work together and get 20 hints & lots of discussion. Rando comes in and sees something they missed, bets yes and solves it themselves.

Or, generally - good guys get a bunch of info, then rando comes in and the good guys have to stop working. Someone can still solve it some other way, thanks to standing on the shoulders of the good guys; and the good guys get nothing. Rando might just never let go of the Yes shares, betting on that. It ends up devolving back to a market that hovers at 50/50 until the hint count is maybe-just-barely enough to solve it, in the market's best guess (which is what I think will happen here)

@Quroe Yes, but you need to gate the biggest yes holders. And they can only hold after we get a bunch of hints. If an agreement like that was made, and I wasn't part of it, I'd go buy yes; which ruins the gains of the agreement.

This is exactly the yudkowsky solution I've kept referring to - except the yudkowsky solution is "Obviously all players would notice this, and would rather solve it and get the 1000 mana than not, even if (as the yes holder) they have to pay some no holders; so we can trust whoever randomly is a yes holder to do this".

But I don't think yudkowsky's solution works in real life (I haven't finished asmodeus yet; but that's the solution proposed at a midpoint in the book)

@DannyqnOht Joshua made a splash in the past by demonstrating the lottery concept worked in this scenario.

To adapt it to this scenario, we'd probably need to say something like, only people who hold [N NO shares] or [have paid X mana for NO shares] are eligible.

@Quroe Sure, that works for knifes-edge bets that can go either way. If you can cause X, then bet on X and then afterwards, go cause X. Lotteries cause X. Great. Joshua was able to capture some kind of gain.

But it doesn't work when betting on X makes X less likely; or when you need to force a market to report that odds are not what they actually are; or when you can't capture gains from causing X.

You still haven't gotten around the fundamental problem here. There's no way to keep out strangers. You can't bet on "ass" if you make this lottery, because someone else will do it first; and you can't beat someone else to it, because then you'd get no hints. You need the random stranger who you can't predict will actually bet yes first, to agree to do the lottery; and we all have to know that'll happen before we start cooperating and betting "no".

I predict this market will get hints until it looks like we have a 51% chance of solving it; at which point the hints will stop until it looks like we have a 49% chance of solving it; back and forth so that we actually will in the end get exactly enough hints to make us on average think we have a 50% chance of solving it given all those hints.

And as such, the true odds for this market are 50% and the liquidity should be trapped and Ben won't lose any mana.

@DannyqnOht We don't have to prevent a random stranger from defecting. We just need to increase the incentive for people to counteract defection and respond in kind.

@DannyqnOht

And as such, the true odds for this market are 50% and the liquidity should be trapped and Ben won't lose any mana.

Oh, haha! You're right! Unless somebody is super sure of their guess, Ben can just close the market before saying an answer is correct and safely collect the uncaptured liquidity. This is risk free to him!

Edit: Risk free-ish

How can anyone counteract a defection? I mean, I literally think that requires some kind of social action against them (eg beating them up).

Anyone can bet yes at any time, and there's nothing anyone can do to "incentivize" people not to do that, or to respond in kind (other than throwing money at the yes better, which they'd happily take, and which the thrower will never get back because there's only 1000M to gain from cooperating here)

@DannyqnOht Respond in kind = bet NO under 50% when somebody bets YES across the 50% threshold.

@Quroe Yeah, that's not really punishing them, is it?

That's just paying them (if a cooperator pushes it back down to "no" so they can buy more "yes". )

@DannyqnOht

Corporate needs you to find the difference between punishment and incentive.

They're the same.

The system would need to reward having an active stake on NO, not just making a NO bet and immediately flipping it for value.

At the end of the day, we simply need to find a way to make the Nash equilibrium sit at 49% to turn the hint spigot on and keep it on.

@Quroe So... All the cooperate people who sign your contract.... Are gonna incentivize strangers to buy "yes"? So the contract signers are gonna work together to pay anyone who to reduce hint chances?

And remember, there's only 1000 m to gain from coordination here. If your contract signers spend 10K mana to keep the price below 50%, the total paid out by the contract-corporation-entity is more than it can gain.

So, the contract can spend at most 999m to keep things at 49%, fighting back the tide of random strangers. And then can expect to gain 1m for all that work.

I think you're missing the point - the core issue here is that useful work done by contract signers (betting no, discussing, planning in public places) can be offset by random strangers benefiting from that work instead of the contract signers. There's no way to punish the random strangers; only to pay them off to get them to stop getting in the way. All strangers have incentive to ask to be paid off. No contract can possibly pay them all off (profitably).

@Quroe I think the nash equilibrium is at about 49% naturally, see my prediction:

https://manifold.markets/bens/will-manifold-solve-my-puzzle#0gy5u6w5sot9

It'll hover at 49% for as long as the market actually thinks we might not solve it, and then it'll go up to 50% until the market thinks "wait actually we need more hints please".

I don't think any coalition can change this without somehow getting enforcement effects on strangers (or else praying that everyone's read Yudkowsky's decision theory stuff, haha yeah right that's not our world)

@DannyqnOht I'm workshopping an idea here. Its not perfect yet. I'm theory crafting. Play ball with me for a bit and hit the idea back to me.

Say, for the sake of argument, I make 1,000 mana limit order on YES at 50%, and it gets fully filled. Cool, I now have an active stake in wanting a YES resolution, and to get there, we need hints.

I make a promise that says, "If this market resolves YES, I will pay one random person who holds at least 100 NO shares at time of resolution 100 mana."

How do you think this will play out? (Genuinely asking, not rhetorical.)