Will Manifold implement Uniswap v3 for liquidity injection?
10
20
100
resolved Oct 5
Resolved
NO

My latest obsession is using the Uniswap v3 liquidity injection mechanism to provide "limit orders" but better: https://uniswap.org/blog/uniswap-v3 Some Discord convo: @SG: My view is that USv3 is a much worse and far more complicated mechanism than limit orders. Getting users to understand liquidity provision is hard enough; getting them to understand liquidity provision that's only opportunistically provided within a given range will be even harder. @JamesGrugett: I also don't understand how it can approximate a limit order. The range order they talk about converts your liquidity in one token to another based on progress in your specified range. For us that would let you convert yes to no shares e.g. between range 20% to 50% if you think the probability should be less than 50%. So, to approximate a limit order of buying M$100 at 20%, you'd mint 100 yes and no shares with M$100, then have a range order to convert your no shares to yes shares in the interval [20%, 100%]. I don't think it's exactly the same thing, and it's way more complicated. @Austin: No actually, you'd probably convert them in the interval [20%, 25%], assuming that you're happy to sell at 25%. There's always a value you're happy to sell at, anyways, that's like a basic principle of economics. This lets you provide two pieces of information: your upper bound for buying, and your lower bound for selling, and that makes it such that the liquidity pool can be much deeper/more efficiently capital allocated, plus providers get paid out better for adding this useful information. It's just perfect all around. How to approximate a "limit buy at 20%" while the current prob is 15% would be to put in your M$100 in the range [20%, 20.01%], plus a rider that removes the liquidity once the market passes 20.01%. It seems kind of silly, sure, but the important point is that it provides a single unified mechanism for liquidity injection of EVERY kind. (The way it might be silly to set up a "prediction market" for a 2-party bet, but actually the prediction market is the generalized form; so too are range orders the generalized form of limit orders) And then the other reason is that you get the ability to draw arbitrary numeric market lines once we have Uv3 implemented. Aka by using 2-category CPMM for numeric (MAX/MIN instead of YES/NO, users can put down their own custom buckets with a range order, instead of needing buckets pre-specified beforehand by the creator. And if we want, we can let them still bet on a normal distribution, by setting the buckets accordingly May 25, 1:35pm: Discord convo was here: https://discord.com/channels/915138780216823849/927671564177133618/979098373955125338

Oct 4, 10:01pm: Resolved by admin (i.e. @SG) due to inactivity.

Get Ṁ200 play money

🏅 Top traders

#NameTotal profit
1Ṁ223
2Ṁ200
3Ṁ49
4Ṁ19
5Ṁ17
Sort by:
predicted NO

See @jack's new market:

predicted NO

@Austin looks like this should resolve

predicted NO
Will this market resolve NO on the 25th?
predicted NO
i.e. if Uniswap is not implemented by then
bought Ṁ12 of NO
Add N limit orders of X dollars with Y spacing above/below/both is much cleaner.
predicted NO
Also Univ3 is for mean reverting pegged stablecoin pools, less applicable to pred markets where people are just market making at wide intervals
Point of clarity, "maniswap v3" is the same as "uniswap v3" for the purposes of this market?
bought Ṁ3 of NO
Austin unilaterally implementing seems fairly likely, and it's good norm to do this. However the odds that users will like it / care / that it will stay seem low. > How to approximate a "limit buy at 20%" while the current prob is 15% would be to put in your M$100 in the range [20%, 20.01%], plus a rider that removes the liquidity once the market passes 20.01%. It's not clear to me this is true. Would all M$100 liquidity be converted to YES in this case after being used? Is that how concentrated liquidity works -- all liqudity is swapped from one to the other once you hit a bound? If so that's interesting.
predicted NO
The idea of users ever "injecting concentrated liquidity" as such seems very unlikely to me. But if you can build multiple different interesting things, in addition to normal limit orders, from this one mechanism, then that seems interesting. I don't really understand why numeric markets would need buckets in any case.
bought Ṁ80 of YES
Okay well even if I don't convince James, there's always the option of me just implementing it and proving it was the right choice. See also: 1-click betting
bought Ṁ30 of YES
I think I'm convincing James!
predicted NO
@Austin you're not haha
bought Ṁ60 of YES
... if only we had limit orders via liquidity injection so I wouldn't have to monitor this market in order to capture more gains.