Stephen Boyd is a renowned @Stanford researcher known for his oft cited textbook, multitude of INFORMS/IEEE awards, and advising BlackRock on convex analysis for manage trillions of dollars
. @akshaykagarwal just defended his PhD under Boyd and is known for his work on visualizations of embeddings via his open-source package PyMDE (minimal distortion embedding)
He’s also a core developer of cvxpy (quant traders ♥️ him) and previously worked on TensorFlow 2
Unlike our other papers, which assume knowledge of crypto, DeFi, + convex analysis, this book chapter is pedagogical + from first principles
Goal: Quant-y undergrads who know Multivariable Calc and Linear Algebra (with proofs, like Lang) *should* be able to pick up CFMM theory
We also show a few new nifty features of CFMMs
1/ Simplified proofs
a. Round-trip trades always lose (path deficiency)
b. Liquidity Provider (LP) share value ∝ ∇ϕ(R)’R [= ϕ(R) for 1-homogeneous functions; surprisingly simple!]
c. Input + output portfolios disjoint
2/ Explicit formulas for add/remove liquidity
Previous papers assumed reserves were constant
We provide a connection between the trading function gradient and the change to liquidity ▶️ helps improve concentrated liquidity formulas (e.g. @Uniswap V3)
e.g. result below:
3/ Exchange Functions
Our curvature paper only showed properties of liquidity (e.g. curvature at fixed reserve) are too state dependent
We elucidate some properties of changes to liquidity via _exchange functions_ which turn out to be concave/convex (*w/o* metric properties)
Their metric properties, which do depend on a particular parametrization and reserve, are shown to be easily computed numerically
This, again, is very useful for measuring impact to concentrated liquidity (e.g. you can extend by linearity exchange functions to piecewise convex)
Finally, exchange functions generalize the invariant calculation done by @CurveFinance to general CFMM curves
There's a simple Newton iteration (gradient descent) for computing trade size from exch. functions
[Remember when @samczsun found a bug in curve's Newton iteration?]
4/ Expected Utility Portfolio
We provide some LP strategies for different utility functions
If we view an LP’s contribution to a CFMM pool as a portfolio allocation, we explicitly find both linear and Markowitz convex programs for how to optimize LP allocation
These are *easily* solved on a laptop and we numerically show how LP allocations change as a function of risk-aversion (cvxpy code included!)
We hope that a clean presentation of these results can make the field more assessable to folks in theoretical CS, ML, statistics, and other quantitative fields
But what’s next? You’ll have to wait until next month ✌🏾
Three items are behind a wall and a solver is going to get one of them for you
Do you get a goat or a car, anon?
@malleshpai, @ks_kulk, @theo_diamandis and I show you that if solvers have to do more work to deliver the item to you, they're not going to show up to the auction
There’s been a lot of talk about `intents’. What are they?
Simply put: they're markets for transaction execution where third parties called solvers compete to satisfy user orders (and any constraints those orders come with)
Question: What are the principal-agent problems here?
tl;dr: We find broad conditions for oligopoly in intent markets
What is oligopoly here? 1. Fewer bidders, k, than the maximum possible, n, participate (i.e. k/n → 0 as n → ∞ or k=o(n)) 2. Users get *worse* prices even though the number of (potential) solvers increases!
While @artgobblers isn't exactly my cup of tea, the novel NFT auction mechanism ford is cool from an auction theory perspective— but is it incentive compatible (IC) for both buyers and sellers?
tl;dr: It is *not* IC but can be modified to be IC!
Quick recap: A gradual dutch auction (GDA) is a sequence of n auctions whose initial prices a_1 < ... < a_n are increasing but where the price of an auction decays as a_i * p(t) where p is non-increasing (e.g. p(t) = exp(-t) in the original paper)
Why would you use such an auction? If you have a series of NFT auctions (e.g. an edition, a daily @nounsdao auction, etc.), you want to incentivize users to pick bundles (e.g. any subset of items to buy) without forcing all of the supply on the market (reducing auction revenue)
1. Winning mechanisms in DeFi come from where you don’t expect (e.g. Uniswap vs. bad copies of TradFi products badly) 2. MEV-aware designs take advantage of FHE/ZK improve efficiency/costs 3. Treasury management isn’t just a meme
The most amazing thing in the filings is how abysmal centralized lenders (incl. exchanges) were at 2 critical functions: 1. Dynamically setting loan-to-values 2. Liquidating bad loans
It’s embarrassing that @gauntletnetwork’s weekly DeFi governance proposals are faster than CeFi
The fact that Voyager never liquidated FTT or SRM and that exchanges were sitting on so much GBTC and *still* didn’t decrease loan-to-values for their largest customers is insanely irresponsible risk management — and it is 100x easier to adjust this as a CeFi lender than in DeFi
Even more embarrassing is the fact that most CeFi lenders didn’t do (as far I can tell) their own liquidations of collateral — they would simply send it to an OTC broker (like Alameda — who probably liquidated their own loans!) and hope things checked out 🫡
One funny thing about DAO governance is that it makes it hard to manage assets held by a collective (why vote to sell the asset you're voting with?) — but for many collectives, asset management is more than pure yield optimization
The goal of Aera is to make it possible for decentralized, censorship resistant treasury management to make it possible for DAOs to hedge their portfolios according to their own KPIs, goals, and objectives while also seeding new protocols with liquidity in a positive sum manner
I've spent a lot of time trying to understand why futarchy didn't work (I'd recommend this @VitalikButerin's post from 2014 for an intro) and tried to figure out if we've learned enough from DeFi over the last 3 years to overcome those failures
Why would a validator agree to this & make off-chain agreements (OCAs) to avoid paying x%?
1. PBS makes this expensive (see @VitalikButerin’s heuristic analysis) 2. Restaking (@eigenlayer) enforce redistribution while reducing variance for validators