The Language of Execution
Vol #133: June 15th, 2023
In order to set up a digital asset fund, the second thing that I bought was an office grade printer.
(The first thing I bought was bitcoin. The first thing I sold was FTX June futures against it, trading with 3.5% of premium over spot and expiring in less than 60 days. Free carry doesn’t last and is never free.)
A printer that could turn out a few dozen pages a minute and scan double sided documents was necessary for two reasons. Obviously there’s a lot of paperwork to complete. But in the halcyon days of early 2021 crypto, there were white papers raining down from every corner of the metaverse.
A couple of pages foretelling paradigm shifting revolutions, supported by a flow chart and the odd differential equation was enough to get a few million dollars of funding and a wallet full of tokens. I’d print these and take them out to lunch on Fridays so I could parse what was opportunity and what was smoke and mirrors.
You don’t need the narrator to tell you what happens next, but in early June of that year, one actually did hit differently. It was over fish and chips at The Dearborn that the implications of the days old Uniswap V3 washed over me like a crisp IPA. (Food has always been a strong memory anchor .)
There’s a not entirely facetious argument that the main purpose of digital assets is to trade digital assets and hope that “numba goes up”. Naturally the early DeFi innovations were around different methods to exchange them. But unlike the halls of the NYSE that are behind two turnstiles and three badge swipes, for a trustless swap to occur there needs to be a different type of orderbook.
The concept of an automated market maker (AMM) was proposed by the godfather of Ethereum Vitalik Buterin back in 2017. Hayden Adams and the Uniswap team made this a reality shortly after, and the design methodology has been copied across every single blockchain.
On traditional exchanges customers looking to trade will source liquidity from the market makers supplying bids and offers. In an AMM, liquidity is provided passively by stakers, who deposit assets to the pool for customers to trade against. It’s not dissimilar from the currency exchange booths that still haunt airport terminals; if the pile of euros is going down because people are trading in dollars, the exchange rate of Euros/Dollar should go up. (Not Financial Advice: Don’t pay that bid/ask spread).
The exchange rate between these two assets in the pool fluctuates according to demand imbalances relative to the size of the pool. If you show up at a Thomas Cook with $1M in cash, not only is the IRS going to ask questions but you’re going to completely drain that booth of euros, skyrocketing the exchange rate. The more money pools behind that euro/dollar pair, the less slippage there is.
Uniswap V3 was a level step in crypto market structure innovation because it allowed for concentrated liquidity. Instead of having your pooled assets fluctuate across the entire spectrum of possible prices, you could concentrate that between say .95 and 1.05 dollars per euro. Greater liquidity in this range meant customers experienced less slippage at the current spot price and the stakers captured a greater share of both the trades and the fees.
It took about a year before the practice was perfected, but clever traders started using the V3 model to provide “Just in Time” liquidity. Since all pending transactions are public in a blockchain (how else does a validator know what to include in their immutable ledger update?) it’s possible to look for large orders, pay a hefty transaction fee to jump to the front of the line, to thus capture all of that large order’s fees with your perfectly deployed concentrated slice of liquidity.
Line jumping aside, it’s starting to sound a lot like a traditional market where the active liquidity providers are constantly shifting bids and offers as inputs change and orderflow arrives. Professional liquidity providers who can be capital efficient will provide tighter markets with less slippage.
They will also start to demand more features. The framework for V4 has just been announced this week and the new functionality here continues to evolve towards the mechanisms that we’re used to back in good old TradFi.
To sum up V4 in one word, it’s about hooks. These are “plugins to customize how pools, swaps, fees, and LP positions interact.” Some of the possible use cases being sketched out range from very TradFi like on-chain limit orders or VWAP algorithms, to the crypto native such as depositing out of the money liquidity into a lending protocol or autocompounded LP fees.
While I was only having my morning coffee when I read about V4, the progression is as natural and obvious as the watershed release of V3. Order types are the language of execution. If you want to build a better place for customers to trade, let them customize their experience.
Execution sits all the way at the end of a trade’s journey. There must be some impetus for why you’re trading. Value investors see stocks as cheap so they buy. Asset managers FOMO herd into the next new thing. Pensioners need groceries so they sell and mutual funds rebalance with bi-weekly inflows.
After doing the hard work of deciding what to buy and for how much, that order now needs to find the other side. There are hundreds of ways to do this. Whatever reason you have for going to the market, this will dictate how you choose to interact.
Different order types allow for different trading experiences. A few weeks back in the Fog of Uncertainty I talked about the delicate balancing act that is execution. Doctors take the Hippocratic Oath, which starts first with do no harm. That’s the absolute best case scenario for most traders. Everyone should have some kind of edge before coming to the market and putting the trade on is mostly about not mucking that up. “No harm” is an A+ for execution.
On a spectrum of trustlessness, one step back from Uniswap is a dark pool. While you need credentials to get in (not very permissionless) the way you interact is anonymous and leaves few artifacts. Further down the line you have the exchanges where you might get tidbits like counterparty information. Sales trading is at the other extreme.
Sales is about talking to people and gaining their trust. While your trade might be guaranteed by the OCC, getting access to the “upstairs” flow in equity options was a people person’s game. It wasn’t just about the box seats; building a relationship with someone through consistent fair dealing meant you would continue to see orderflow. There’s no way to get a guy back on the blockchain for helping out on a tough order.
To be larger than life on a trading floor is big shoes to fill. It takes the kind of guy that drives 200 miles to buy an Audi A4 from a crooked preacher in Pennsylvania with a bounced check and ends up with a fake title. Who else could invent an order type called Cancel If Touched?
It often feels like as soon as you go to lift an offer or hit a bid, whoof, it disappears. Maybe that’s because someone beat you to it - the reasons you decided to trade now are probably similar to theirs. But sometimes it just feels like someone knew you were coming and faded. There are of course many rules against this and audit trails in triplicate to prove it, but what trader isn’t a bit of a conspiracy theorist?
Giving someone a Cancel If Touched order is a joke. But it’s not said without a shade of a wink. A piece of code or exchange matching engine will automatically match buyers and sellers when the price is right, that’s exactly what they’re designed to do. But each side always wants to tease out a little bit more, so perhaps if someone showed interest I’ll act coy. Playing hard to get down at the CBOE. In the slow motion world of sales trading, the order type is a human.
Non standardized markets take the order type customization even further. Real estate markets are notoriously opaque, with shadow listings and transactions that don’t hit the tape until they’re months stale. Offers1 on homes include contingencies like limit price (perhaps with an escalation clause), but also qualifications for mortgages or inspections. Sellers will accept but only pending their own purchases and refuse to part with the dining room chandelier.
Part of the reason a barrel of oil or DOGE are so tradeable is because every unit is fungible - not so with 1830s colonials. The order type in the housing market is the dynamic relationship you have with your agent, their network, and the sellers. To borrow Uniswap’s term, there could be hooks on everything from the septic field to the moon phase.
Options trading started with a guy who had to walk around putting deposits on olive oil presses, and has exploded in complexity with the power of technology. Algorithms can read every bit of market data and react with updated orders in real time, but there’s still a spot for the VIX pit to fill 50 cent on his 100k order for upside calls that would have spooked an electronic order book only a fraction of that size.
Order types are ultimately about navigating under the constraints of certain preferences. Bespoke deals require detailed negotiations about non standard specifics, while large scale trading demands nuanced price slicing in commoditized instruments. Trading is where buyers meet sellers. How they talk to each other matters.
This term has frustrated me ever since I started working in markets. A buyer makes a bid, a seller makes an offer. The prospective buyer should make a BID on a house. All my semantics has gotten me is sideways looks.