We’ve currently got 3 rewards mechanisms available for positive usage of the Rally Network. With this post I’ll attempt to detail the current state of each and work towards a proposal to iterate forward on each of these existing pieces. Additionally, I’ll attempt to initiate a discussion on a new reward system that encourages desirable behaviors within the network. I’ll try to do this in a way that individual design changes explored could make sense as either standalone proposals or combined into a single proposal substantially overhauling the system.
Yield Delegating Vaults (YDV)
YDVs wrap deposits into underlying Yearn vaults. As these deposits earn via Yearn strategies, accumulated earnings are transferred to the community treasury and an amount of $RLY proportional to this transfer is made available as a reward for depositors. The amount of $RLY made available is multiple of the tokens transferred to the treasury configurable on each YDV. The multiple is expressed relative to the tokens transferred to the treasury (i.e. the $USDC vault transfers $yUSDC to the treasury and the multiple expresses how many $RLY should be emitted for each $yUSDC transferred). Since each token type transferred by each YDV to the treasury has a different value, this multiple is different for each YDV and is intended to be relative to the value of the token that YDV supports.
- 500MM $RLY allocated for distribution via YDVs at launch
- Minimize risk by transferring from the 500MM pool to the relevant YDV smart contracts on a periodic basis
- ~30MM transferred to contract to date
- ~20MM emitted as rewards for depositors
- Multiples described above initially set with the intention of doubling the APY of underlying Yearn vaults
- Multiples have not been changed since initial launch configuration
- Price changes to $RLY mean these vaults are currently offering ~200% APY
- Early confusion about reward timing led to further incentivizing the most popular YDV with a per block reward via the Liquidity Provider reward program
- Basic usage data
- Tokens worth approximately $200K remain in legacy YDVs earning no $RLY rewards
- Total deposits in active YDVs ~$26MM
- Recent surge in deposits after holding fairly stable at ~$12MM for the last 2 months
- Currently emitting ~600K $RLY / day while collecting ~14K USD equivalent tokens for the community treasury
- Virtually all deposits are in the 3Crv YDV
- Yearn is in the process of deploying Vaults v2
- Currently unclear what this means for existing Yearn vaults
- Gated launch of v2 vaults showing promising early results with stable coin yields against USDC and DAI topping 30% APY
It seems to me that the YDVs are roughly doing what we’d like them to. It is unclear whether the per block incremental reward for the 3Crv YDV is a substantial driver of its success relative to other vaults with similar underlying returns or if it’s simply the relative attractiveness that goes along with being the biggest.
I’ve seen some discussion around concern that YDV rate of emissions is theoretically uncapped within the 500MM allocated; but I’m of the belief that, since these rewards help build the community treasury, the overall rate of emission is less important that ensuring the returns are attractive enough to motivate deposits and moderate enough that the community treasury is building appropriately relative to the $RLY emitted.
With the above in mind, I’d propose:
- No adjustment to the per block reward granted to the 3Crv YDV from the liquidity rewards program
- Ongoing, monthly reset of the multiplier on each YDV to take into account $RLY price change with a target of 2x return of underlying Yearn vaults
- Tentatively suggest using trailing 7 (or 30d) average price as a basis for resetting multiplier
- Assess compatibility with Yearn v2 Vaults; consider upgrades as necessary
350MM $RLY earmarked at launch as incentives for deposits in the liquidity rewards program.
- Liquidity rewards contract emits 38 $RLY / block (~7.5MM $RLY / month)
- Current pool weights divide these 38 $RLY across 6 pools as follows:
- 13% BAL RLY/USDC 90–10
- 13% BAL RLY/USDC 10–90
- 13% BAL RLY/ETH 90–10
- 13% BAL RLY/ETH 10–90
- 35% UNI RLY/ETH
- 13% y3Crv Yield Delegating Vault (aka curve.fi/3pool LP)
- Community discussion has raised questions about the ongoing incentives for the y3Crv YDV and the relative risk/return for the various liquidity pools against each other and the YDVs
- No clear proposals have emerged but I think the general sentiment conveyed is Uniswap LPs deserve a greater return and the y3Crv YDV share of the rewards would be better allocated back to one of the liquidity pools
The nice thing in my mind about per block rewards is that it should enable pretty direct discovery of user preferences. If rewards seem high, more money will flow; if rewards seem low, less money will flow. At this point, the APYs seem to have settled into a place where depositors are not moving around much and are finding some balance between re-depositing rewards for compounding gains and selling them on the open market. Additionally, the total depth of the liquidity pools relative to daily volume still appears on the high end relative to other comparably sized Balancer and Uniswap pools.
With the above in mind, I’d propose:
- No change to current liquidity rewards
Community Activity Rewards (CAR)
The 7.5B $RLY allocated to promote network usage represents 50% of the fully diluted supply; CAR is the first program that draws from this bucket and currently distributes ~
Details on the design are available here:
Real time view of weekly rewards available here:
- At the current creator count, the total pool made available for CAR is 1000 $RLY / creator / day = 39K $RLY / day
- Rewards are calculated/accrue hourly and are distributed weekly
- The difference in accrual and distribution means all community members holding coin at the time of rewards distribution are rewarded evenly and potentially independent of their contribution to rewards accrual
- Rewarding at the community level makes the program easier to understand for short term buyers/sellers and harder to understand for long term holders
- In particular, it’s not clear to creators how they should think about or value these rewards apart from them generally being good for everyone participating in their community
- Because holders would only perceive the rewards effect through changes in the value of their holdings, there is little/no perception of a positive reward effect if net negative demand for a coin in a given week is larger than the rewards distribution (i.e. if 10K $RLY are added to a coin’s liquidity and 10K $RLY are distributed to users who convert out of this coin, a single holder will see no change)
- Relatively inactive communities create opportunities to redirect large reward distributions to coins with very little real usage to the benefit of the individuals choosing to move in and out of these coins solely to farm rewards
I believe the current reward design began with the thinking that consistently increasing total liquidity backing a coin is one of the strongest indicators of the health of a creator coin community and should therefore be the basis of a reward system that benefits the entire community. However, the mechanics of the calculation and distribution mean we’re not quite incentivizing positive movement in this indicator; instead, we’re enabling active traders to capture rewards independent of their effect on this metric, and we’re generally increasing/concentrating the volatility of a creator coin around the timing of reward distributions.
I believe a straightforward solution to keep this system consistent with original intent would be to:
- Programmatically add and remove coins from CAR eligibility based on a calculation that encompasses usage and active participants. Coins below eligibility thresholds would not receive any rewards for a period
- Distribute coin rewards to users based on pro rata ownership of the coin as opposed to simply increasing the backing liquidity pool
- Align calculation and distribution intervals to ensure anyone benefiting from a reward distribution is doing so in more direct proportion to their impact on reward allocation
- Target hourly but introduce some jitter/randomness
These changes would:
- Prevent farming of inactive coins to the detriment of active communities
- Make rewards distributions more tangible with visible changes to coin balances while
generally preserving the magnitude and equitable distribution of the current system
This would be particularly apparent to creators themselves relative to current state
This should make it easier to understand how rewards work and enable communities to better activate against reward earning behavior
Discourage concentration of activity around a less frequent reward distribution
Other Rewards Considerations
In addition to the above, community discussion has suggested considering the following for either a modified allocation of CAR or new rewards mechanism:
- Sell activity within a community increasing CAR allocation to the benefit of users who remain
- Transactional activity aligned with goods and services moving within a creator economy
- Net platform usage/users attributable to a particular creator coin
- Collaborative efforts spanning multiple creators and/or communities
- Novel use cases pioneered by creators and/or communities
I’d like to invite discussion in this area with the intention of getting to a proposal that allocates additional $RLY from the network usage bucket to a new reward mechanism that exists alongside YDVs, Liquidity Rewards, and CAR. Ideally, a new incentive structure alongside these that introduces incentives roughly on the order of magnitude of the existing CAR or Liquidity Rewards systems focused on desirable behaviors within the side chain creator economies.
I’ve offered my thoughts on the current state along with proposals to evolve the existing rewards systems and begin discussion on introducing a new one. Let’s have a lively discussion that gets us to a proposal for appropriate evolutionary design of our existing systems and the addition of a new reward mechanic that incentivizes important behaviors not covered by the existing system.