Recommend market structures that incentivize betting on long term markets
Ṁ2,500 / 5000
bounty left

Sometimes people don't bet on long term markets because they don't want to tie up their mana for a long time. Put in some suggestions for how to avoid this and I can award a bounty if I like them. Maybe @SirSalty will add extra to my bounty.

From the mana supply announcement:

We are currently discussing other solutions to incentivise trading on longer-term markets. Please let us know your ideas!

Get
Ṁ1,000
and
S1.00
Sort by:
+Ṁ300

If shares held were to pay interest, there would need to be some extra mechanism to specifically prioritize long term holdings rather than short term ones. Otherwise, traders could still hold most of their net worth in shares in short term markets and accrue the same interest as those in long term markets, removing any incentive to specifically take on long term holdings.

Some possible mechanisms:

  • Shares only pay out interest at the end of a fixed interval, e.g. every 3 months. This way short term holdings never accrue interest.

    • Issues with this: whatever interval is chosen would be the maximum term incentivized. There would be no reason to bet on e.g. a market resolving in 6 months over one resolving in 3 months.

  • Shares only pay out interest on specific markets, such as those with a resolution date more than 6 months away.

    • Issues with this: as above, the shortest term market in this category would still be the most incentivized.

  • Have a higher yield on shares the further out the resolution date is (logarithmic yield curve)..

    • Issues with this: markets with max yield, i.e. very far out resolution times, would almost entirely become Keynesian Beauty Contests, as they would mostly attract whales seeking high-yield interest who will want to exit their position at some point before the far-out resolution date. The most profitable traders would be those who best guess the short term sentiment of whales rather than those who best make long term predictions.

    • Would probably get abused by markets with a chance of taking ages to resolve but will likely resolve much sooner (e.g. "Will [public figure] do Y before they die?" where YES is highly likely).

  • Allow traders to specifically purchase interest yielding shares that cannot be sold until the end of a fixed term, with a higher yield for longer terms.

    • Issues with this: prevents traders from changing their position as the market or the underlying facts change, increasing inertia.

    • Could be abused as above, e.g. buying a 2 year share on a market likely resolving this month, getting the highest interest but with a shorter term. This could be guarded against by a variety of mechanisms:

      • The shares don't payout until they vest, so a 2 year share won't payout until the 2 years are up, even if the market has resolved.

      • The excess interest (compared to yield on a term matching the time actually held) plus a tax is deducted from the trader's balance on resolution. The tax is necessary to disincentivize always taking the longest term possible, since doing so effectively grants a loan of the excess interest until resolution. So this option creates default risk.

  • As above, but the shares can be traded to other users.

    • Issues with this: a lot of UI/UX work to implement, only partially addresses first issue.

  • Allow traders to specifically purchase interest yielding shares at the cost of extra transaction fees, such that they aren't profitable over regular shares until they've been held for e.g. 3 months.

    • No obvious issues or avenues for gaming the system I can think of.

    • Naturally creates a logarithmic yield curve as the transaction fee is a sunk cost that is less and less significant the longer you hold for.

    • May still require a few tiers (higher transaction fee = higher yield and longer term until profitability) to properly incentivize various term lengths.

+Ṁ50

Dividends: Interest on long-term bets that accumulates in the form of shares to all shareholders. (With a multiplier factor on the age/duration of the bet, tuneable)

+

Currency demurrage: tax on uninvested balances (averaged over time period, area-under-the-curve, not at gameable points in time)

=

Balanced inflationary and deflationary forces. Set the dividends to exactly equal the demurrage.

+Ṁ50

Keep loans and limit loans to say the 90% VaR. This means you don't give someone a loan if "based on the market probabilities, there's more than a 10% chance that this person will go bankrupt".

Pros:

  • I could write code to do this for you guys

  • This keeps the current loan system which is popular and results in good calibration of long term markets

  • This prevents irresponsible lending which will cause a lot of users to go bust and leave the platform. You can use an even higher VaR than 90% depending on what risk you want to allow users to take.

  • How much a user can borrow will depend on how much risk they have taken in their portfolio. Someone will all their money on a 50-50 question won't be able to borrow anything, but someone with their money diversified across a lot of 90% questions could borrow more.

Cons:

  • Manifold has said they want to do away with loans.

  • It's a system that's hard for people to understand, but if you tell people their 90% VaR everytime they get a loan, people will get an intuition for how much more they can borrow and how a bet will affect that.

  • Might be computationally expensive to implement

+Ṁ50

Fee dependance on the bet length.

When making a bet I choose, how much the invested mana is locked before I can sell it. Longer I say, less fees taken. If I click "up until resolution", then the fee for the transaction is the minimal value possible.

Long markets thus are a tiny bit more attractable, because if fees are not imposed, then your bet and thus profit is bigger.

If you bet on a "resolves in 24h market", then the fee can be huge.

+Ṁ50

After thinking about how not getting loans will change my betting... I'll still want to bet on long term markets because I'll want to track them, will want to know what the answers will turn out to be, and will want to know how I did. But since they're a terrible "investment" relative to shorter-term markets... I'll just bet tiny amounts of mana. All the benefits of participating in long term markets, without the costs in terms of mana lock-in. To me it doesn't really matter if I made 100 units of fake money or 10 units of fake money or a thousand units of fake money, so betting ten and doubling my "money" 5 years from now feels about the same as betting 100 and doubling my "money" 5 years from now. But betting 10 means I can make more bets without running out of fake currency, given my current bank balance, so that's what I'll do. When the loan system was giving me x% back per day, interest free, I had in mind that continued participation would allow me to build up a significant amount of money for charitable donations over several years, but that isn't going to happen now - but I'll still play for the "see how well you can predict future events" fun.

Make it so that people like me are catered for :D such as:

  • If a market is very long term and the bet is very small, give some of the mana back after a time period, so I can bet again, if I'm still active in Manifold. This restores the incentive to continue betting in long term markets, without massively inflating the mana supply, and will let me participate in all the long term markets I want to without worrying very much about running out of mana. As long as I can check in on a long term market once a month and bet in it when my credence changes because of new information, I'm happy.

  • Let bettors take notes of what their credences were when they made the bet, and what their rationale was. My tiny bets aren't going to move the market, so I won't be able to go back in a year and say "I bet the market up to X, so that's probably how likely I thought it was, at the time", but being able to put a note in about why I made a particular bet fixes this.

This isn't going to get you good long term predictions, but it'll keep people who aren't interested in gambling in the "closing soon" markets, but also aren't going to buy a bunch of mana so as to move markets that close years from now, interested in participating.

+Ṁ50

Replace the fee with a monthly wealth tax of 1% of your net worth - all the value of the bets that you have had for over a year. This both incentivises long term trading and removes the fee. Removing the fee also incentivises trading on all bets

+Ṁ50
+Ṁ50

Another option should be to make sure that the value of mana is stable/ to have deflation by simply not giving out sny loans and doing other various deflationary things so the value of all investments will naturally grow in real terms. This will also make simply having mana profitable but the profit you could get on long term investment are still a nice plus.

+Ṁ50

Manifold's central bank could keep lending to users but at some interest (that could be changed to manage mana supply) and some part / all this interest could be payed as a form of dividend to users who had a bet on something for over 6 months in proportion to the value of their shares. This does not create mana supply problems as the Manifold central bank could charge a fee before handing over the interest to long term shareholders but it still makes both leveraged betting and long term betting possible.

+Ṁ50

One possibility is to create a chain of derivative markets, in which each one is resolved based on the value of the next one.

Suppose we want to create a market for an event in far future. We create a series of markets, each one running for 2 years. The market M[I] is created in the year X+I (where X is the data when the question was originally asked) and is resolved in the year X+I+2 based on the value of the market M[I+1]. If the final resolution happens in the year Y, then the market M[Y-X-2] is resolved based on the actual event, and no further markets are created.

+Ṁ50

Have different currencies for different time lengths. Mana for 1week long markets/predictions, Energy for 1 month long, Chi for 1 year long, Essense for 2 years old and so on (arcana, aether, vitality, prana, elixir). Those currencies cannot be converted. Each has its own leaderboard, but the main leaderboard uses multiplication of all the score of a person. You would need to perform well in all leaderboards to get a nice multiplied value.

+Ṁ50

Another option would be to simply keep the loan system but only allow users to take 1% in loans every day and only after having kept shares for over 6 months.

+Ṁ50

Have a Manifold ran bank lend money to the users at some interest defined by supply and demand whilst the supply is the pool of the long term bets (or some part of it). If the pool is then given back that interest (minus an eventual fee), it will grow in value making betting on long term markets useful.

+Ṁ50

Markets which resolve N/A shortly, but which will be re-resolved to the correct resolution in the future.

For example, say you want to know when AGI is out. You set up a numeric market which closes by, say, at the end of this month. When it closes, the market is N/Aed and all mana is returned. When AGI comes out, the market is re-resolved to the correct date.

Or say you want to know whether US will send aid to Ukraine before 2024 ends. You set up a normal market, but which resolves N/A at the end of this month. After 2024 ends, you re-resolve the market to the correct outcome.

You can't profit by buying based on market sentiment, and there's not much opportunity cost involved.

+Ṁ50

The equivalent of "door prizes" awarded periodically to people in proportion to how long they have been in a market and how much mana they have invested. So even if your mana appears to be locked up, it might not be, you might get rewarded for holding a position for a longer time.

A minor alternative: drizzle in bets by distance to close.

reposted

Now that the real money aspect is public, people should start punching in more ideas after taking that into account. This bounty still has a lot of juice in it, even if it is worth 10x less than it was when we started....

Well. Ahem. Devaluing the currency by 10x with a few days notice is one thing that will NOT encourage long term bets.

It's one thing to think "the mana flow at this rate cannot last, but they might find some way to maintain the value of bets people have already made, in whatever solution they come up with, many things are possible and you have no history to go on", and another to think "Manifold may change abruptly at any time, and past history indicates long term bets are not protected from massive devaluation".

I'm here partly because good betting can lead to charitable donations, so I'm going to be selling 90-99% of most of my positions (leaving pennies in to indicate bet direction and track questions) to make what charitable donations I can before the exchange rate changes.

EDIT: Sold what I could for a disappointing amount. The loan system continues incentivising some of my long term bets, because I owe more in loans than the positions are currently worth.

Edit again: I see that they are willing to work with those with large balances to not devalue the donations they may have been planning to make (which is the only thing mana can currently convert to) which means my criticism above should be much less intense. When writing what I wrote earlier, I was a bit sad because "hey, I 10xed my balance in 4-5 months, from $10 to $100 USD - if I can do that again this really starts to add up" was a large chunk of what was motivating me to continue to participate, last week, and now, like, nope, I'm almost literally back to exactly square one, when I got $10 USD of mana to play with. But after calming down, I think the management is doing what they can in the most impactful cases, and I knew things could take a left turn at any point when I signed up, I'm not out any money, and my time was spent learning things, so, shrug,

A "happy hour/day/period" at the beginning of each month (when people have lots of free mana because of markets that have recently resolved) where fees on markets that don't resolve for a while are reduced or eliminated.

More generally, you want the price structure of participating in markets to reflect demand, where long term markets are cheaper to participate in than short term markets.