ParaBoost Retrospective - Analysing boost usage and effectiveness after 20 epochs

Research and Discussion
1

Two years ago, a new experimental overhaul was passed to the tokenomics system as PSP 2.0. The proposal aimed to address some of the shortcomings of the initial staking system and introduce two new experiments in staking that were not commonplace at the time: liquidity-based staking and action-based boosts.

Recently, a community thread began discussing a revamp of the existing distribution system, with different ideas being discussed. Laita Labs supports discussing new incentive innovations, and considering that a common complaint of the modern staking system is its complexity, and believe we must first assess what parts can be streamlined before introducing new ones.

With over 20 distribution epochs since the new system was introduced, we can now assess how many stakers have been engaging with the boosts, their commitment over time, and the overall revenue performance as an incentive to accomplish these metrics.

Liquidity & Stakers Assessment

Since the introduction of the sePSP2 token (A balancer pool token that serves for voting and staking), the amount of non-PoL TVL locked in the 80/20 Balancer pool has increased significantly from ~$3M at the time of the introduction of PSP 2.0 to peaks of $18M. At no point has liquidity been lower than before the introduction of PSP 2.0.

In addition to helping boost liquidity, sePSP2 has also been found to be an attractive staking option compared to its single-sided alternative. By the end of epoch 20, the TVL for sePSP2 was $4.8M compared to $1.15M in sePSP1.

PSP 2.0 was also effective in increasing the number of unique addresses staking PSP in Mainnet and Optimism. From PSP 2.0 first epoch until epoch 20, the number of stakers increased 167% from 1,843 to 4,927.

ParaBoost Assessment

Action-based boosts - aka ParaBoosts - were introduced to drive higher volume, feature utilization, and staker engagement. Thus, we can assess the success of the system by asking ourselves three questions about the boosts, ordered from highest to least important:

  1. How much growth in volume did the ParaBoosts bring to the ParaSwap protocol?
  2. Did the ParaBoost lead to an activation of features within ParaSwap stakers?
  3. Has the ParaBoost system usage remained consistent throughout its deployment?

While some of the ParaBoosts led to an activation of a minority of stakers, none of them managed to fulfill the objectives established.

Currently, the DAO incentives are using 80% of all the revenue generated every month, all while requiring time consuming and error-prone complex calculations to execute this system. The latter issue in particular has created considerable challenges for the automation and analytics of the distribution process.

With all of this in mind, the next sections present to the community the Data which has driven our decision to reconsider our approach for the long-term efficiency of our ecosystem.

Trading Boost

Due to the main function of the ParaSwap protocol, the trading boost is among the most important introduced in PSP 2.0, as it incentivizes consistent use of the protocol by stakers across epochs compared to passive stakers.

Regarding utilization, an average of 12% of stakers used the stable swap boost, while 37% used the volatile boost. Overall, the volume generated by these boosts represented less than 8% of the total ParaSwap volume.

In addition, the fees generated by these boosts represent less than 4% of the total ParaSwap protocol fees. The number of stakers using these boosts and the volume and fees generated by them have decreased consistently since the introduction of PSP 2.0.

Chart1572Ă—972 41.9 KB

Chart1572Ă—972 49 KB

Chart1600Ă—991 81.7 KB

Chart1600Ă—991 93.4 KB

Making Boost

Considering the prior market-maker-based system of PSP 1.0, the Making boost served as an extension of the trading boost for RFQ makers. However, after comparing the usage between these two, we can begin to see the divergence between these two types of trades.

On average, less than 1% of stakers used the making boost. While the volume generated by the Maker boost represented less than 0.9% of ParaSwap’s total volume.

Chart1504Ă—928 40.5 KB

Chart1504Ă—928 41.9 KB

Referral Boost

By accessing the protocol through referral codes, both the referrer and the one referred to receive a portion of the fees generated during that swap.

To incentivize community-led onboarding, the initial proposal suggested also introducing a boost for existing stakers. Users could receive a ParaBoost to their existing stake by using referral links and referring someone.

Across all epochs, volumes generated by referral trading represent less than 1% of the total volume generated by ParaSwap, all while being used by less than 2% of stakers. Additionally, since the introduction of the system, it has been consistently decreasing.

The lack of utilization suggests that the incentives driven by the ParaBoosts were not sufficient to mobilize the use of this system. As such, it could we believe it might achieve better activation by introducing a separate system in the future dedicated to referrals only.

Chart1200Ă—742 26.4 KB

Chart1496Ă—926 30.1 KB

Holding Boost

With the holding boost, users are incentivized to keep part of their stake liquid in any chain, boosting users who didn’t stake their entire PSP amount. On average, less than 7% of stakers used the Holding boost.

Chart1200Ă—742 31.3 KB

Pooling Boost

With the introduction of a Balancer-pool-based token, sePSP2 introduced a staking option that would unlock incentives to stakers while still enabling liquidity for the protocol. The objective of the Pooling Boost was to incentivize long-term staking of this option in significant amounts, as to stabilize the long-term sustainability of the protocol and enable long-term stakers.

As there is a minimum requirement of 60k PSP staked in sePSP2 to apply for this boost, looking at the overall usage by stakers doesn’t make sense for this boost. Instead, analyzing epoch 20 data, we observe that 86% of stakers using the boost have been staking for 6 epochs or more, meaning they reached the Pooling Boost cap, indicating that the Pooling incentives might’ve assisted in long-term stability.

Chart1200Ă—742 25.4 KB

Summaries and Proposed Next Steps

After examining the usage of all boosts in the ParaBoost system, we can draw the following conclusions on how effective the different incentives were:

  • Introducing a liquidity-based token, sePSP2, has unlocked a level of liquidity depth higher than at any point before its introduction.

  • Only a mere 8% of the overall ParaSwap volume and 4% of the overall ParaSwap Fees were generated by the most widely utilized boost, the Trading Boost, which showed an average of 37% of stakers engaging with it.

  • The making and referral boosts were both used by a negligible number of stakers, and their impact was not significant to their initial objectives

  • 86% of stakers using the Pooling boost have staked for 6 epochs or more, meaning that thanks to it, the amount staked in the system remained stable across all epochs

We believe that the following next steps could help advance the ParaBoost system:

  1. Simplification of boost system: With some of the boosts not being utilised by the vast majority of the stakers, these serve only to add complexity to the system and hinder on staker adoption. Additionally, it is blocking the way for new incentives to be added to the system.

  2. Automation and verification considerations: Considering the amount of time and complexity needed to calculate each boost, the system’s few benefits are outweighed by its challenges and the manpower needed to run it. Additionally, by streamlining the system, we permit not only easier external verification of the rewards but also simpler (and perhaps automated) execution and verification of future automation and integrations.

If simplified, the staking system could focus on incentivizing the most successful part, the establishment of a permanent liquidity base, while other parts can be spun out to separate proposals. Having said this, we propose creating a future proposal to remove the Trading, Making, Referral, and Holding boosts and leave the Pooling Boost and the 2.5x sePSP2 boosts untouched.

We would like to thank all the community members who have engaged in feedback and discussion with Laita about what could drive the next wave of adoption for the ParaSwap protocol. We would like to invite everyone again to this thread to discuss the next steps of our DAO’s mechanisms.

7 Likes
4

I would vote for it.

1 Like
5

I find the Trading Boost beneficial as it incentivizes all of us, holders, to use the protocol that we are supposed to believe in.

I strongly agree that other boosts are mostly useless, as you demonstrated.

I would support a redesign or removal of the boost system.

1 Like
6

We could cut 10% (or more?) from the normal distributing stakers fees and use them for some developpement: security vault, marketing, any improvement. And the best, in this case would be a flexible allocation. Something we could vote for - as PSP stakers - each 3 months or 6 months (time as to be determinate). For exemple, after an analyse provided by someone from the team (these are the people most aware of the needs of the protocol from my point of view), and after a discussion and a vote, we could choose where this allocation goes for the next 3 months or 6 months or whatever we would choose. Something useful for long-term protocol development.
This is an idea that awaits improvement from the PSP community.
This would encourage the community to participate in gouvernance.
Please let me know what you think about this idea.
Thank you

1 Like
7

The idea is good, however I think that the operation of the pooling boost should be reviewed, by modifying the 60,000 psp level and perhaps opting for a calculation that would take into account the number of psp with the number of epochs.

For example with random numbers.

Already put a limit of 100% of parraboost in the boost pooling.

To reach it, put a threshold of 100,000 psp in the sepsp2 with an increase of 10% per epoch. The increase of 10% per epoch being the maximum. This would be a linear system. A person with 40,000 psp in sepsp2 would increase by 4% per epoch.

This would allow small wallets to be able to benefit from this increase in parraboost on their scale without being limited by the 60,000 psp level. And the biggest wallets would have an advantage without abusing this system because they are blocked at 10% by epoch.

This would favor the holding of sepsp2 in the long term.

3 Likes
8

First of all, I would like to thank the rest of the Laita team for taking the time to analyse the boosts and provide the graphs, this helped a lot to bring things into perspective. The trading boost making up less than 4% of fees (with the latest one being even less than a percent) was a big surprise.

Considering the technical and accessibility debt that this complex system is causing, I think it’s time to Grab the Chainsaw to free up the system and allow new things to grow from there :seedling:

Regardless of what the new system entails, I feel that most important lessons going forward is to focus on flexible and composable systems. Our current system relies on the Governance committee for creation and verification of proposals because of all the nuances it has. If we are able to incentivise the system in a nearly automatic manner, that will allow parameters modifications to allow the DAO to have a more direct voice in the future!

2 Likes
9

We share that the current ParaBoost system is too complex and difficult to understand, and the metrics shared are eloquent in demonstrating that such complexity generates very little impact both in terms of user usage and volume for the protocol. We have no observations nor do we see undesirable consequences in simplifying the system in Pooling Boost, which has a high adoption/usage rate, but on the contrary, the proposed simplification of the system will not only benefit the protocol by reducing management efforts of underutilized mechanisms and facilitate future automation for greater efficiency, but will facilitate understanding by stakers.

In the community there is concern about modifying the fee and/or token distribution system of the PSP token, and/or establishing a burn or buyback, so it is logically desirable that before addressing this issue, it should first address its simplification, so as not to spend efforts in modifying ParaBoost mechanisms that are proposed to be deprecated.

Based on all of the above, we support this initiative and we are excited to see the feedback from stakers and the community.

What would the process of automating executions consist of? Any ideas in mind? Is there anything we in the DAO can collaborate on?

10

As @Laita’s post shows, based on the metrics provided, the Pooling Boost has been successful in terms of getting PSP holders to keep their tokens in stake, so we find this idea, which deepens the concept of the boost being linked to the lock time, thus lowering the barrier to entry into the system, very interesting.

We understand that the numbers given are random, but keep in mind that we are talking about a PSP threshold of 100k, which is equivalent to $1.5k, being 10% just $150. But regardless of that, don’t you think that putting a threshold on the boost could be a disincentive for stakers who are currently contributing more than the threshold? We have to be very careful not to have the opposite effect to what we are looking for and end up creating a disincentive to stake.

11

To help long term holders paraboost, why not move to a time based system with epochs. This would also ensure that everyone starts on an equal footing, regardless of the amount of sepsp2.

On the basis of always keeping SEPSP1 in X1 and SEPSP2 in X2,5.

I use random numbers, for example we could decide on 4 levels of blocking possibilities for SEPSP2 and SEPSP1 while keeping the possibility of not having a block on several epochs :

  • Level 1 giving an additional +25% but if the staker wants to withdraw, they would need to wait 5 epochs before getting their tokens back.
  • Level 2 giving an additional +50% but if the staker wants to withdraw, they would need to wait 10 epochs before getting their tokens back.
  • Level 3 giving an additional +100% but if the staker wants to withdraw, they would need to wait 15 epochs before getting their tokens back.
  • Level 4 giving an additional +200% but if the staker wants to withdraw, they would need to wait 20 epochs before getting their tokens back.

For example a wallet unlocking its SEPSP2 with level 4, a reduction in paraboost will apply per level until the SEPSP2 are fully unlocked.

We could possibly implement a fast unlocking system that allows for faster withdrawal but with a compromise of returning SEPSP2 to other stakers. For example :

  • Level 4 could ask to exit at the next epoch by slashing 80% of their SEPSP2 to the stakers.
  • Level 3 could ask to exit at the next epoch by slashing 60% of their SEPSP2 to the stakers.
  • Level 2 could ask to exit at the next epoch by slashing 40% of their SEPSP2 to the stakers.
  • Level 1 could ask to exit at the next epoch by slashing 20% of their SEPSP2 to the stakers.

Those who would leave early would therefore give an additional incentive to those who want to stay by obtaining an additional yield in SEPSP2 and SEPSP1.

This is a draft and the numbers for epoch per level, paraboost percentage and slashing percentage are to be adjusted.

I think we need to move towards a simple system of understanding regarding revenue sharing for as many people as possible, while leaving the possibility of benefiting long-term stakers.

12

Proposal to Enhance Long-Term Staking Incentives for ParaSwap

Introduction

ParaSwap, currently offers two staking options that impact voting power and revenue:

  • sePSP1: 1x multiplier for voting power and revenues.
  • sePSP2: 2.5x multiplier for voting power and revenues, with an 80% PSP and 20% ETH pool.

These staking options reward users for participating in the protocol. However, the current system lacks strong incentives for long-term staking. This proposal introduces a new mechanism to incentivize long-term holding of sePSP, while ensuring the system remains fair and beneficial to loyal participants.

Please note that all figures provided in this proposal are examples only. These parameters can be flexibly adjusted by the protocol to fit the system’s needs. The formulas governing these values are straightforward and can be found at the end of this proposal.

Please feel free to give your opinion after reading this. Feedbacks are heavily appreciated.

Proposed Changes

1. Locking Period and Paraboost

Users can choose to lock their sePSP tokens for a period of up to X epochs (where X is the maximum, and configurable by the protocol). The longer the lock, the higher the paraboost applied to both revenues and voting power, rewarding long-term stakers.

I believe the paraboost should also apply to voting power, as long-term stakers have a vested interest in the protocol’s success and are more likely to contribute to its improvement over time.

  • No Lock, No Boost: Users who do not lock their tokens will not receive any paraboost.

  • Sliding Lock System: To enhance user experience, a slider in the interface will allow participants to select their preferred number of epochs, ranging from a minimum to a maximum. As users adjust the slider, they will be able to see the corresponding paraboost value grow dynamically in real time. This ensures that users can easily understand the impact of their chosen lock-in period on both their revenues and voting power.

While the exact number of epochs can be adjusted, as an example, locking for the full X epochs could yield a significant boost of up to 167%. Users are free to lock for shorter durations, but the paraboost will be correspondingly lower.

At the end of the user-selected locking period, the paraboost returns to 0%. Users will need to relock their tokens to maintain any paraboost benefits, which encourages continuous long-term staking commitment.

In the following sections, I will take a pre-determined maximum number of epochs for the boost of 26 epochs which corresponds to two years, and a locked up period of 13 epochs for the staker. Therefore the staker would get a 63% paraboost for a year while still being able to withdraw its stake with a slashing penalty.

With this system, here would be the ParaBoost compared to the number of epochs where the stake would be lock.

Locked Number (Epochs) Paraboost during the duration of the lock
1 4%
2 8%
3 12%
4 16%
5 21%
6 25%
7 30%
8 35%
9 40%
10 46%
11 51%
12 57%
13 63%
14 70%
15 76%
16 83%
17 90%
18 97%
19 105%
20 113%
21 121%
22 129%
23 138%
24 147%
25 157%
26 167%

2. Withdrawal Flexibility with Penalty (Slashing)

While users can withdraw their tokens before the lock-up period ends, they will face a slashing penalty based on the remaining epochs. The slashing begins at 60% and reduces epoch after epoch in a non linear manner.
This system not only incentivizes users to lock their tokens for a higher number of epochs, resulting in a greater ParaBoost, but also discourages early withdrawals. With each passing epoch, the slashing penalty decreases, making it more advantageous to wait for the next epoch rather than unstaking too early and losing more of their stake.

Here is a table so you can better understand this system (it changes according to the duration of the lock, but the slashing always begins at ~60%) :

Epochs passed before complete unstaking process Stake Slashed Decrease in the Slashing compared to last epoch
1 59% -
2 56% -3.16%
3 52% -3.40%
4 49% -3.67%
5 45% -3.95%
6 40% -4.25%
7 36% -4.58%
8 31% -4.93%
9 26% -5.31%
10 20% -5.72%
11 14% -6.16%
12 7% -6.63%
13 0% -7.14%

3. Slashing Redistribution

When a staker withdraws before their locked period ends, a portion of their tokens is slashed as a penalty. These slashed tokens are not burned; instead, they are redistributed proportionally to the remaining stakers in the protocol.

This system is designed to reward long-term holders, who believe in the protocol and maintain their positions, while penalizing those who exit early. In essence, this redistribution reflects the natural order: stakers who stay committed are compensated for their patience, while short-term holders sacrifice a portion of their tokens. This mechanism ensures that long-term stakers continuously benefit from the impatience of others, increasing their revenues over time and thus their holdings.

4. Lock-in Starting in the Next Epoch

To prevent opportunistic behavior, the lock-in period will only begin in the next epoch. This ensures that no one can stake immediately after observing slashing events and benefit unfairly. During the current epoch, users can still withdraw without penalties, but they will also not earn any rewards. This keeps the system fair and prevents gaming of on-chain events.

Comparison with ZeroLend and Curve

The proposed system introduces significant improvements over protocols like Curve and ZeroLend:

  • ZeroLend: Locks $ZERO for up to 4 years, and diminushes the revenue and voting power of each $veZERO accordingly and linearly to the period locked.
  • Curve: Works similarly to ZeroLend, here is their working mechanism :

veCRV is an acronym for vote-escrowed CRV. Users can lock their CRV for a minimum of 1 week, maximum of 4 years, in return the user is given veCRV, veCRV amount decays linearly over the chosen lock time. veCRV is not transferrable. The longer you lock the more veCRV you receive, see the locking formula section for a detailed explanation but the simple explanation is:

  • 1 CRV locked for 4 years = 1 veCRV
  • 1 CRV locked for 3 years = 0.75 veCRV
  • 1 CRV locked for 2 years = 0.5 veCRV
  • 1 CRV locked for 1 year = 0.25 veCRV
  • …

Our proposal, in contrast, offers a much more dynamic and rewarding model. With our system, locking tokens longer results in exponentially higher rewards (for both voting power and revenues), and users have the flexibility to withdraw early, albeit with penalties.

Appendix: Formulas and Explanation

1. Paraboost Formula

The paraboost formula rewards long-term stakers exponentially.
Paraboost = (1 + 1/X)^Y
Where X is the maximum number of epochs and Y is the number of epochs locked. This formula ensures that as the number of epochs increases, the boost grows exponentially, providing strong incentives for long-term staking.

2. Slashing Penalty Formula

The slashing penalty is designed to penalize early withdrawals while protecting the long-term stakers.
Stake Slashed = 1/(Z^k)
Where Z is the maximum multiplier, and k is the number of remaining epochs. As the stake duration progresses, the percentage of the stake slashed decreases, encouraging users to remain staked longer.

Conclusion

By implementing this proposal, ParaSwap will foster long-term staking, create stronger incentives for loyal participants, and establish a fairer system for both revenues and governance. This approach not only enhances user experience but also positions ParaSwap as a leading platform with innovative and user-centric tokenomics.

1 Like
13

Hey @Cubionix , thank you for suggesting an alternative to the current staking system!
Could you share it in a separate research post? This way the two topics (simplification of the incentive system and introduction of an alternative) do not get conflated

1 Like
16

Currently, all operational processes related to the incentives system are tied to data processing by the laita team, followed by internal verification by the governance committee.

We believe that in the future, we should explore processes that improve the transparency and automation of the distribution system. This could range from anything like onchain execution of the proposal to onboarding a service provider to handle distributions automatically, just like mimic does.

By simplifying the system, this kind of execution will become possible, and provides an additional layer of safety, as it is not only less prone to human error, but also more efficient compared to manually creating and collecting signatures.

1 Like
17

Thanks for your feedback! Just to clarify, I’m currently working on modifying the existing vision for the paraboost system and the pooling boost you’re talking about.

Since these changes build on the current proposal, making a separate proposal would conflict with the ongoing improvements and might create confusion.

My goal is to fine-tune the existing system rather than introduce an entirely new one.

18

We find the approach of this draft very interesting and we encourage further deepening, but we have to keep in mind that Balancer is about to release its v3, which will include the possibility to use and create hooks to extend the functionalities.

Therefore, we believe that we should first focus our efforts on removing the current complexity of the ParaBoost system, as suggested in the first post (remove the Trading, Making, Referral and Holding boosts and leave the Pooling Boost and the 2. 5x sePSP2 boosts untouched) to simplify the system and make it as simple as possible until the release of Balancer v3, as this is the right time to migrate to the latest version that includes the hooks when they are implemented, having infinite possibilities to think, use and create mechanisms to suit the needs of ParaSwap.

So we understand that it would not be opportune to create a whole new veToken system from scratch now, when Balancer is about to release its v3, which will extend the possibilities that exist today. Let’s simplify the system today and when v3 is released let’s think about how much more possibilities are open to ParaSwap than today.

We even think it would be positive to contact someone from Balancer to deepen these topics and get professional advice on the new version.

2 Likes
19

I am

I totally agree with that.

Plus i would add that the simpliest/easiest solutions are often the best ones.
Watch the staking solution offered by AAVE. It is simple and it works.
You stake your assets to add more security to the protocol and earn Safety Incentives. In the case of a shortfall event, your stake can be slashed to cover the deficit, providing an additional layer of protection for the protocol. This vision brings confidence, because you know where you go.
Plus you have one, and only one cool down period. Easy.
This is, from my point of view, the direction we should take.

3 Likes
20

Thank you everyone for the discussions shared on this thread.

We have shared a formal proposal to move forward with the simplification, which you can find here: PIP-XX: Streamlining of the staked PSP incentive system

As @SEEDGov has noted above, this proposal will focus first on removing the complexities of the system. In the meantime, we look forward to reading on research and discussion regarding potential improvements that could follow afterwards.

4 Likes
21

Makes perfect sense ! ready to vote on it - bring it to snapshot frens :rocket:

2 Likes