https://manifold.markets/LivInTheLookingGlass/will-donald-trump-be-indicted-for-a

Loan of m90, expected value of 440 YES shares at 31% is m136.

I think the odds are lower than 31%, as the year is running out, but I think I get more mana back if I stay invested and wait for more loans. So I think I will keep the YES shares as hedges against betting NO on other markets.

Which I think resolves this market YES unless I'm missing something.

I don't think I'm at the loan cap, my profile still says I'm not getting loan income.

I found an old market with a loan of m20, value of m1. LTV around 2000%.

https://manifold.markets/GustavoLacerda/will-russia-invade-kiev-by-end-of-2#Cp5JJpb1N2BgClttmHh1

I think the odds of a new Kyiv invasion are now below 3%. If I expect the resolver to not count the battle for Kyiv then I should sell yes. But loans means that is -m19 cash flow. Since there's a high chance that the market will not resolve, it's better not to sell. Seems pretty clear cut.

@MattP Any thoughts on this?

@Yev Points out that this is 2,000% LTV and the market says 50%.

Also @MartinRandall, worth noting the question you're actually asking would be better formulated as discouraged from "day trading" on a market, as this only really applies to repeated buying and selling. When I initially saw this market I assumed you were just talking about any trading at all - and this "problem" only arises when dealing with day trading for marginal profits (and again I go back to: the point of the loans is to incentivize long trading on long markets, it isn't a flaw if they don't also incentivize day trading).

One thing I'm curious about - do loans come due if you buy NO shares (to cancel out) rather than selling your YES shares?

@MattP yes they do, James is not that silly.

I mean, I did note this to him, so who knows whether he would have not taken it into account otherwise :P

@MattP Would you expect a market "Will a public coinflip per year produce the 5th head before the 5th tail?" to become more or less accurate due to these loans?

@Gurkenglas That one's easy - more accurate, because before them no one would bother betting at all on such a long term market and after them there's now enough incentive for new people to jump in and correct the market when it becomes mispriced, even if some people already in the market don't want to do so because they'd rather keep mispriced shares than pay the loan back earlier.

I'm no longer confident Martin will be convinced - I can see how the (short term) negative cashflow from redeeming a loan early to get a bit of extra EV might discourage some traders from doing so. I am confident this is still a positive change for long term market accuracy overall. If y'all don't want to correct obviously mispriced long term markets you already hold loans in, there are others willing to do it (and more willing than before there were loans).

@MattP I don't think day trading is the right term, because the bets need not take place over a single day. I can certainly update the description to try to clarify.

If I find that I am not discouraged because when I am thinking about selling with high LTV I never expect to subsequently buy, or that seems unlikely relative to the profit I am making by selling, that would resolve this market no.

What's the concern here exactly? The way it works, the loan on any future bets is not affected by past bets unless you're at the cap.

If you bet M$100 on a new market, 5% of that bet is returned week 1, 5% of M$95 week 2, etc.

If you previously had a M$500 bet in that market before making that M$100 bet, well... 5% of M$100 is returned week 1, 5% of M$95 week 2, etc. This is now in addition to whatever loan schedule you had with the previous bet ofc. I'm not seeing the disincentive, maybe you can explain it to me?

Feel free to bet NO then. A worked example:
- I have a 2000 YES shares, worth m1000, on a 50% market with a m500 loan.
- I believe the true probability is 50%.
- The market probability increases to 60%. I sell my 2000 YES at 55% (after fees and spread) for m1100, or m600 after loan repayment.
- The market probability decreases to 40%. I buy 2000 YES at 45% (after fees and spread) for m900.
- I now have 2000 YES shares, worth m1000, on a 50% market with no loan.
- My expected profit is +200, but my cashflow is -300. I have also spent time analyzing the market and placing bets.
- The -300 cashflow reduces the net present value of placing the above trades, discouraging me from trading on the market.

@MattP Ignoring the cap, if you have 2000Y bought for M$1000 on a 10-year market because you expect to sell them for M$1200 in a year, and a year hence you could only sell them for M$800, then without loans (or fees) you want to sell if you no longer thing you have any information, and with current loans you'd lose M$130 by liquidating your position.

If the M$800 price is EV-neutral, you'd be incentivized to stay in the market. Perhaps you'd put in a limit order to sell at >M$930, then comment to the guy who originally sold you those 2000Y for M$1000 ("correctly" expecting to buy them back for M$800 in a year) that it would be worse than inaction for you to sell below M$930.

To balance the incentives, we might let you find some other person in a similar position as yourself, such that both of you like the other's position more than their own, and let the both of you swap positions while keeping loans. This platform is supposed to be Manifold Markets after all. But then we might as well introduce a medium of exchange to that barter economy, and let you keep the loan if you sell only to keep betting on long-term markets. I've previously suggested having multiple currencies of different time scales.

@MartinRandall in this example, you're still better off cashflow-wise (at worst, even) in all cases than you were before the loans were implemented. Work the same example, but assuming you never got a loan at any point.

You're also not counting in this example the cashflow from buying those shares initially - adding that in would correctly show you as negative cashflow in both end cases (which makes sense, because you're holding an asset that matures later), but either slightly less negative or significantly better off in all cases than you'd be if there were no loans.

@MattP Your argument of being better off than before the change proves too much: It would also endorse multiplying everyone's mana by ten.

@Gurkenglas why do y'all keep saying you'd "lose" money by paying back the loan when you sell the shares? You already gained that money in the past! You are better off than if you'd never gained it in the past! I feel like I'm taking crazy pills here.

It's better to have M$500 for a while to mess around with and then pay it back at no interest than to not have that M$500 at all. If you incorrectly formulate the problem as only including the repayment of the loan, but not the original disbursement of the loan, obviously it looks bad. Stop doing that - that's the wrong way to formulate the problem!

@MattP Let's raise the stakes, I've posted a limit order at the 60% you seem willing to trade past.

@MartinRandall Re: cashflow, see if working that same example starting from the proper beginning (aka from the first purchase of shares, through all loans, through all trades, looking at cashflow at each step), with and without the loan scheme, changes your mind. If not, I'll sell my NO shares because I ain't got much better than the hard math to try to change your mind and ultimately, this question is about your subjective feeling.

@MattP He should do your calculation when considering whether to initially buy in. He should do our calculation when considering whether to go through with the sale. This temporal inconsistency may well converge to a norm that the buyer and seller of 2000Y, who both expect +20% within a year, agree to settle their bet according to their new expectations a year hence, splintering our markets.

@MattP I think my decision to initially buy in is a sunk cost / sunk benefit when it comes time to decide whether or not to sell later on.

Do you disagree? Am I missing something?