Hello there dear numerai collegues,
did you hear about SingularityDAO? I would like to start discussion on this project with relation to numerai.
Short description of SDAO: It’s a spin off from SingularityNET and its purpose is to be decentralized hedge fund. They start with dynamic crypto sets, but want to expand into stocks later. The beta version of dynasets is starting right now and anybody can come and stake their stable coins, eth or singularity assets to earn or burn interest from the trading models.
The questions I would like to answer are:
* Is it “numerai killer”?
* How is so that anybody can stake on dynasets and nobody can stake on numerai meta model? Is it because of the crypto market vs the stock market?
* Is numerai team in contact with singularityDAO team? I think numerai and singularity are both amazing projects that tries to do the similar thing with different approach, maybe cooperation could benefit all.
* Why numerai tournament never went into crypto? If crypto is here to stay, isn’t crypto the future stock market?
* What are your thoughts?
So this is kind of complicated. To really answer your question someone would need to be pretty knowledgeable on both projects and finance, to properly distinguish the differences. Without that basis, almost every answer is going to be strictly an opinion, and that kind of lends itself to the circular “yeah but” type responses.
At the end of the day, the projects may share some broad similarities, but the differences will likely be numerous and quite deeply ingrained in each firm.
It’s always easy to ask “why”, but it can be hard, if not impossible to answer. Numerai, for better or worse, is a ship that has set sail with a specific goal in mind and they’ve built a company around that end-point. After an unfairly brief review of Singularity, it doesn’t appear to me they are currently headed into the same waters as Numerai. I think the day a pension fund is willing to invest $150m into a crypto fund is way off into the future. So, Singularity may be the fund of some distant tomorrow, but I think Numerai has the most potential in the near-term, for more traditional institutional investors.
The killer could be Numerai like performance but that allows the participants to invest/stake as well. Or where there is some form of feedback between the performance of the fund and the token.
The above example is based on what is called a “long only” strategy. DynaSets may be run using multiple types of strategies including “short only”, “long/short”, “neutral” and also variations where our DAM automatically swaps between strategies based on market conditions.
For investors like pension funds and fund-of-funds, style-drift isn’t really a feature as it completely disrupts any risk modeling they’ve done. I think the target market here for the foreseeable future is going to be retail, and transitioning beyond that could be very difficult (the backtest article is problematic).
This isn’t to say Singularity doesn’t solve a problem for individual investors in the crypto space, but it’s not the same problem Numerai is tackling on the equity side with regards to HF inefficiency using ML. Right now any comparison is pretty much apples and oranges.
That makes sense. Their evaluation function is also different. They are trying to beat the average percentage gain of underlying assets; thus, they don’t mind to have 20% loss if hodlers holding over the same period of time have 21% loss. If you are hodler you will use their service. So yes, I would say that what numerai is doing is harder, and what SDAO is doing is more sexy for retail.
Before I went to bed yesterday I was thinking about the problematic volatility of NMR. Yes volatility is nice if it goes up; however, the current trend of NMR is at most underwhelming. I think that the lending option mentioned in the other thread would be great solution, but why stop there? What if pple could put NMR in a NMR dynaset, get a token similar to LP tokens and stake those at the tournament? If the only reasons for staking at the models are: to have skin in the game, and to determine which models are more trust worthy, then the potential risk of the NMR dynaset should not matter right? Am I missing something?
The idea I had in my mind was similar to how LP tokens work. If you lend a token to a pool, to provide liquidity, you get an LP token in exchange. 1 LP token basically says: “You own 1 NMR at the pool”. Thus, IMHO it should be possible to create something similar with dynaset, something like DS token. There would be an option in model setup (like it is with MMC and CORR) that your staked NMR would be in the dynaset. When you stake a NMR it would go to the dynaset and you would get a DS token instead. If numerai wants to burn your NMR, they would take away your DS token, and burn the NMR later. As I understand it, it does not matter whether a value of DS token match a value of NMR, because the reason of burns is to limit bad actors and speculators with random models.
Maybe there is a simpler solution how to do it. The example should have been proof of concept. I don’t know how others, but I would love to have an additional alpha to my stake.
@sneaky finding new use cases for NMR would be great! At the moment, even returns from the best models are probably going backwards when the value of NMR is taken into account. That can change of course, but I don’t think anyone would complain if there were more options to reduce the crypto risks or add alpha or yields.
Though whatever the idea, it has to be compatible with the main function of NMR. Numerai are in the business of running a hedge fund and probably have not excess resources or interest to manage a dynaset. So this idea needs more work to figure out how it would work and if it is feasible or not.
But hey, keep turning over those stones and digging deeper, you might be getting close to finding something!