This is a derivative of
The TL;DR is that I think Manifold should test out having some markets where no one is allowed to spend more than 100 mana, or perhaps hold more than some number of shares. I don't think most markets would be improved by this cap, but I think some could be.
This market is intended to gauge public sentiment on not just if they will try this, but if we will think it was a good decision to do. It uses a shorter timeframe, ending in July of 2024.
If Manifold implements this idea, even as a test, I will put up a poll after a few weeks of implementation and ask if it was a good decision to implement. The poll will run for one week, with the options "yes", "no", and "see results". I will then resolve this market according to the results of the poll.
A problem with this market, is that we will not know if it was good idea before a long time, we will only know if it drives probabilities toward something that seems more reasonable (ie : if it drives it towards our own credence).
And surely, it will drives these probabilities, on average, towards the credence of most people, because it drives the mechanism toward a vote system.
I am skeptical of the idea, I am unsure how to justify it, would it really lead to better probability if bets are capped ?
It would change nothing if users in a market were not interested enough to bet a lot (or not enough rich to do that), and it would drive the probability toward a vote system when most would like to bet more than the cap.
And if it is a mix, it restrains the power of the more interested/richer users.
I am unsure this is good, and also it seems to me, like it is still not a well justified change, and I feel like it doesn’t address the core problems of long term markets.
One problem is the loan and league system, and as I say in my other message, I think it is fixable in another way.
An easy way to show realized profit (or even better, current maximum profit if you sell), and choose it as the default graph to show on the home page, would also help with the incentives.
But it seems to me that one of the core of the problem, is that the users betting incorrectly and a lot on long terms markets will not lose their mana before a long time (and users betting correctly will not earn a lot of mana before long time).
So the only thing we have, to induce who is good or bad at long term predictions, is who is good at short term predictions.
I copy past your answer there, to answer to it here.
But these long-term markets really do seem like a problem to me, because most people will look at them as the same kind of forecast as all the other markets on manifold when really there's a complex system of incentives that arise from Leagues, the Loan System, and the chances that Manifold ceases to exist before those markets can ever pay out.
I think the loan system is incorrect, I understand why they changed it, but they introduced a new problem by doing it.
I think a rule should be : You can’t increase you loan by more than %4 of what you bet, on betting on a market. That is, if you have 1000 shares on a market at 10%, and you bet 10 on it, increasing the probability to 20%, your loan should not increase to 4%*1000, but to <=4%*10.
So you can’t bet on a market to increase your loan more than how much you bet.
A similar rule should be used for leagues.
Unsure what is the best way to do it, the old loan system have this property, but also the problem that your loan don’t decrease when the probability decrease. I think "how much mana you can earn by selling" would be the correct way to do it (then if other people increase or decrease the probability, your loan still change).