Sense Just Got Better for Everybody

An update on what Sense achieved in Q2

As we’ve discussed in our Why Sense article, we believe that the logical solution for self sustaining, permissionless, decentralized, resilient fixed income infrastructure is yield stripping. Yield stripping is the only way to provide safety (full collateralization of fixed income), abstractability (can make fixed income equivalent of any yield-bearing token, including rewards), and flexibility (can divide value between principal and yield tokens multiple ways, and use those the fixed income equivalents in other DeFi applications). Yet, no yield stripping fixed income protocol has seemed to achieve product market fit.

Why?

Sense yield stripping contracts create a permissionless market with four types of participants. In order to get the flywheel of market action, every market participant has to be able to get what they need out of the market. For Sense’s original design (as well as every other yield stripping protocol), there are blockers to the specific value proposition sought by each of those market participants. Our goal this quarter was to identify and solve those core value prop blockers. Below we’ll discuss what each of those blockers are, and how we solved them.

The four groups that must come together for Sense are liquidity providers (LPs), fixed income buyers (PT Buyers), variable yield risk buyers (YT Buyers), and on-chain actors that call protocol maintenance functions (Protocol Operators). Below is a diagram of the “Sense Flywheel”: a self sustaining market between all needed parties. We include each of these four actors, and what the core value prop of participating in the market is for each.

Let’s discuss each of these actors, what has been blocking their core value proposition, and how we fixed it last quarter.

Liquidity Providers

Liquidity providers to yield stripping protocols want to get an attractive return profile on a passive “set it and forget it” position, with a higher expected value than just keeping the provided assets for themselves. Before Sense, frankly, they didn’t. The issues really boiled down to two problems:

(1) Being an LP for yield stripping protocols had a bad return

(2) Being an LP for yield stripping protocols required active management

Let’s talk about the root of these two blockers to the core value proposition for LPs and how Sense is fixing them.

The Unattractive Return Profile of Being an LP

The best AMM solution for trading a fixed rate product before Sense created the Space AMM was a traditional Yieldspace Invariant AMM. The Yieldspace Invariant AMM was a brilliant solution to the impermanent loss that would otherwise be caused (in a constant product AMM like Uniswap V2 or Sushiswap) by the value drift of the fixed rate token vs its underlying: a fixed rate ETH product is expected to appreciate over time in ETH terms.

The problem is that when you are providing liquidity for that fixed rate asset in a traditional Yieldspace Invariant AMM, you are dividing your capital between that fixed rate asset and a non-yield-bearing quote asset. Let’s use an example from Sense.

First, you have what we call a “target asset,” which is some variable yield-bearing token or vault: let’s use wstETH, which provides a variable staking yield from Lido’s ETH staking network. Then there’s the “underlying,” which in this case would be stETH. Through the process of yield stripping, the protocol has also created a fixed rate version of wstETH called sP-wstETH, which we call the principal token (PT).

In a traditional Yieldspace AMM, an LP would provide liquidity in terms of sP-wstETH and ETH. This means, a huge chunk of their capital is getting no yield! Unless trading fees (or some temporary external incentive like liquidity mining rewards) compensates the LP for providing liquidity, they’d probably get a better deal just holding sP-wstETH, or even the target asset wstETH.

Obviously a lot goes into how returns play out, but here is a bit of a scenario analysis. In the below chart, we look at the value of a share in a traditional YieldSpace invariant pool over the course of a year in seven illustrative cases. In every example, the starting priced in fixed rate for sP-wstETH (the fixed income principal token) is 10%, and for the sake of simplicity wstETH yields are flat at 10% and trading fees are 0. Each line shows the performance of the Yieldspace Invariant LP share vs just holding wstETH given a different path for priced in implied rates for sP-wstETH.

In the majority of cases, the fact that a large portion of your LP capital is earning no yield causes you to underperform just holding the target asset instead. The one notable extreme exception is when the implied rate (IR) goes extremely high: people are continuously selling sP-wstETH into the pool until the pool hold almost all sP-wstETH that is trading at a large discount to the ETH you can redeem it for at maturity (even right before maturity). In any more realistic case, unless you have large trading fees or external incentives, you expect to underperform just staying in the initial asset: wstETH.

Basically, the expected value of being a liquidity provider vs a benchmark like wstETH or your fixed rate sP-wstETH is negative.

Why would you choose to provide liquidity to the pool above instead of just holding wstETH? You probably wouldn’t, unless you were farming an airdrop.

How Sense Fixes This

Sense fixes this with the Space AMM, which is a logically simple but technically difficult solution. Basically, instead of using non-yield-bearing ETH as your quote asset in the AMM (which dilutes your yield), Space uses wstETH (or whatever the target asset it), such that even with no trading, all of your capital is accruing yield in your Space LP position.

This makes the Space AMM LP outperform the traditional Yieldspace Invariant LP in any case where rates are not negative.

It would be impossible to design a pool that in every case outperforms target, but Sense Space’s extra yield makes it an attractive return profile vs target, even excluding any trading fees or external incentives. This means LPs actually get a good deal–a positive expected value vs holding wstETH or sP-wstETH. As you can see below, there is some risk of underperforming the target asset (wstETH) depending on implied rate price action, but the expected value is better: in more cases the LP share outperforms.

For more detail on the Space Pool LP return analysis referenced above, check out this deep dive.

The Pain of Having To Actively Manage a “Passive” LP Position

One of the things about having fixed income tokens with set maturities is, well, they mature. This means that being an LP isn’t actually passive. If you’re in, say, the 5 year sP-wstETH Space Pool this isn’t such a big deal. But, if you’re trying to provide liquidity for a monthly fixed income token, you have to move your liquidity into the new 1-month token every month.

That is, frankly, annoying (and can have tax consequences).

How Sense Fixes This

Sense’s solution for this is, again, technically tricky but logically simple: auto-rolling LP positions. Basically, you can provide liquidity for a 1-month (or whatever duration you want) and as soon as it matures, it will issue a new 1-month automatically, and roll your liquidity into the new one (with a configurable “cool down” period where users can withdraw their funds with zero slippage before setting the new pool up).

At time of writing this is in audit, and will be released soon. See our announcement of the feature for more details.

Variable Yield Risk Buyers

Variable yield risk buyers want to get outsized returns in a capital efficient way for taking risky positions on variable yields. But before Sense, they actually couldn’t.

The appeal of YTs is that they theoretically provide really capital efficient exposure to rates: if I convert 1 ETH into 1 ETH worth of YTs while the implied rate is 1% for a 1 year maturity, and then implied rates go to 2%, my position is worth 2 ETH. The value of my position changes in “yield space” — the % change with the previous implied rate as a base (i.e. if the rate goes up from 1% to 2% it has risen 100%, not 1%, and the value of my position has gone up correspondingly).

Unfortunately, it’s that bit about “if I convert 1 ETH into 1 ETH worth of YTs” that has historically been a problem.

Buying YTs Isn’t Worth It

Before Sense buying YTs from a yield stripping protocol was also a pretty bad deal. Here’s the root of the problem with buying and selling YTs: if there is a 1 year sP-wstETH maturity that is currently trading at a 5% implied rate, and I mint YTs and PTs with 1 ETH worth of wstETH, I get 0.95 ETH worth of PTs and 0.05 ETH worth of YTs.

So how do I mint 1 ETH worth of YTs? I either need to:

1. Keep running that transaction over and over again: sell the 0.95 ETH of PTs back into ETH, then mint another 0.95*.05 ETH worth of YTs and repeat. This would mean my face getting ripped off by gas fees.
2. Use 20 ETH to mint 20*.05 = 1 ETH worth of YTs. This is horrendously capital inefficient, and gets even sillier at low durations: I would need something in ballpark of 240 ETH to do the same thing for a 1 month series. “CaPiTal EfFiCiEnT,” eh?

Even when providing a secondary market for these so you can just “buy” the YTs, you run into problems: the market cap of YTs is way, way lower than PTs, and their value is volatile (bad for AMMs), and goes to 0 after maturity (also bad for AMMS–but note, here, that depends on whether you have streaming payments to YT holders vs only post-maturity payments). Basically: minting YTs sucks, and buying YTs also sucks.

How Sense Fixes This

With the release of our “Efficient YT Purchase” feature Sense solved this issue. Again: technically tricky, but logically simple. If you have 1 ETH and want 1 ETH worth of YTs as in the example above, on the Sense portal you will in a single transaction borrow the other 19 ETH you need from Sense (as long as there’s enough ETH in that adapter), mint the YT using 20 ETH, and pay the 19 ETH back to Sense with no interest.

This means, for the first time, YT buyers can actually access the core value prop of “I can use my 1 ETH to take a 1 ETH bet on yields.” The capital efficiency of YTs is no longer theoretical: it is real.

Fixed Income Buyers

As described above, it has (until now) been unattractive to buy the risky “YT” side of a yield market from a yield stripping protocol. How does this affect fixed income buyers? It results in lower yields.

Fixed income buyers want fixed yields rather than variable yields, for financial planning or asset liability matching, and should at some point be willing to take a lower than expected average yield to offload the variability risk. The higher the demand for the risk on the other side of the trade, in the form of YTs, the higher the fixed yield PT buyers can get.

In other words, more people buy YTs, the higher the fixed rates on PTs.

If buying YTs sucks (as described above), no one buys YTs, and yields are smushed (or there is very little liquidity for yields at a given level).

This is really just an extension of the solution above, so I’ll leave the discussion here at that.

Protocol Operators

The need for protocol operators to be incentivized is unique to Sense vs other yield stripping protocols, as existing yield stripping alternatives were not designed from the get-go to be permissionless. As it stands right now, Sense is the sponsor of all existing series, but incentives are already built in such that self-motivated on-chain actors can profitably take action (and, in fact, have taken action) to maintain the protocol. We call these operations protocol incentivized ops.

And, as it turns out, it works!

How Sense Incentivizes Permissionless Maintenance

Sense is built to incentivize protocol operations that can be done permissionlessly by any on-chain actor.

To give an illustrative example: a deposit that was put in by the sponsor of a series at its inception, plus any issuance fees collected by that series, get paid out to whomever calls the function that “settles” the series at maturity. The initial sponsor of the series gets a grace period to settle and reclaim their deposit + those fees (collectively called the “settlement reward”) before it becomes an open function anyone can call.

For our first ever matured series Sense didn’t call the function to settle the series we had sponsored within the grace period where we had unique access to it as sponsors.

Immediately upon the grace period ending, an MEV bot came and settled it for us. The bot paid $582 in gas (it was mid May) to collect a deposit worth $668.27 and collected issuance fees worth $53.71 for a tidy profit of $139.47.

You can find the transaction here.

What it All Means

We at Sense believe that fixed rates are a critical, missing part of DeFi infrastructure. We also believe that yield stripping is the correct way to approach this, and that it can be built to operate permissionlessly.

At the beginning of this past quarter we looked around and asked ourselves why product market fit and adoption for this kind of asset had lagged behind other on-chain applications like money markets and DEXs.

We realized that the core value props for market participants were being blocked by failings in the design. We identified each of the blockers to those core value propositions and made it our goal to remove them. And we did!

Now that most of the solutions are in production, we’re focusing more on adding additional assets, TVL growth, etc, for the current quarter.

If you’re excited about this vision and want to help build the future of finance, join us! We’re actively hiring and are always on the lookout for top talent. Join our community on Discord to get involved, and follow us on Twitter for updates!

Further reading:

To use Sense to LP in Space Pools, buy fixed rates, or speculate on yields: Sense Portal

To learn more about Sense’s vision: Why Sense

To learn more about the return profile of being and LP: Space Pool LP Returns Analysis

To learn more about yield stripping protocols and how they work: Sense Core Concepts

To learn more about auto-rollers and how they work: Auto-Roller Announcement