How will Manifold loans work at the end of March?

Plus

Ṁ1488

resolved Apr 29

100%38%

Exactly as first implemented

1.8%Other

As first implemented, with a different loan amount

1.3%

No longer automatic

0.5%

Removed entirely from Manifold

The same, but with an option to not use them

With significant differences such a positive interest rate

14%

Loan limit improves with your performance to boost long-term accuracy.

1.5%

Add the ability to pay real money to increase your loan amount.

29%

Markets with resolve dates farther in the future are eligible for larger loans

0.9%

High probability bets (eg YES on a 95% market) are eligible for larger loans

On March 1st, we implemented the ability to automatically take out interest-free loans on the first M$ 20 on each market, inspired by a @ScottAlexander proposal https://astralcodexten.substack.com/p/play-money-and-reputation-systems There's been some community discussion over the mechanics of this; this market is a place to express your thoughts on which direction to take this feature.

This question is managed and resolved by Manifold.

Get

1,000

and

3.00

18 Comments

23 Holders

86 Trades

Sort by:

Why don't I see the option for a loan

This

Yeah thanks for the reminder!

This should be resolved before we all forget what the situation was on April 1st

Seems like it could get weird when someone closes a market early.

This might not be the smart money, but this is what I would like to see (or something like this) so will chuck 20 at it.

I misunderstood how the loan system worked, causing me to be overly pessimistic about it.

> In other words, now my earnings become a function of how many markets I invest in rather than just the quality of my bets. This makes Manifold have the same problem as Metaculus. Not really. In Metaculus you can select answers that will give you income, no matter how market will be resolved. On Manifold you still need to be above average predictor to get profits.

I totally agree with the principle mentioned in the recent newsletter post: > First, this makes it more attractive to invest in long-term markets. Before loans existed, it was hard to justify putting your money into a market which wouldn't resolve for a while. Even if you were very confident that a 50% market would resolve to YES in 10 years, doubling your money over 10 years only represents a ~7% yearly rate of growth. It'd be much easier (and more psychologically rewarding) to make that much on markets resolving in weeks or days. It seems like a straightforward extension of this would be to make the loan amount dependent on the "close date" for the market, i.e. users betting on longer-term markets get larger loans than users betting on short-term markets.

I'm not a fan of the loan system, although I appreciate both the motivation and the desire to experiment. For me, it totally removes my ability to track how well I'm predicting as a function of growth in terms of my principal because I can go and bet $20 on 50 markets and now have as much money invested from loans as I had originally. In other words, now my earnings become a function of how many markets I invest in rather than just the quality of my bets. This makes Manifold have the same problem as Metaculus.

Having the loans as default worsens the "smart money predict[s] what the average person will bet a loan on, and/or convince[s] them to do so" problem. Someone who has to decide to strategically switch to a loan is putting at least that much thought into their bets... unless they're REALLY bad and running entirely in the negative XD

In a sense making your loan cap depend on the total money you'd have if you sell everything would do this too. And it's very close to how loans work in reality.

(Also, this is the kind of gamification that for-profit companies like, right?)

The loans system gets long-term markets moving at all, which is good. Currently it makes smart money predict what the average person will bet a loan on, and/or convince them to do so. What you want is for smart money to predict and cause smart bets, so you want loan limits dependent on money smartness.

Having automatic zero-interest loans takes the fun and the stakes out of it for me. If portfolio maximization is the only goal, the best strategy is to take that M$ 20 loan and invest in each and every market (which is not *fun*). Am I missing something here?

Positive interest rate would be complicating things a lot, and make harder to adjust distorted markets. And difference would be tiny if present at all at 20M$

*such as a

Interesting! Can you speak about why you'd be interested in the option not to take a loan? (also: is this option different from "no longer automatic"? I'm thinking this option is "loans are default on", while "no longer automatic" means "loans are default off")

Related questions

Related questions