Hi everyone.
Happy to join the conversation as has been requested of me - I appreciate the request by several of you privately for my honest input. I have taken quite some time from when I first asked until this post so I have had enough time to reflect and talk to various types of parties- this is a very nuanced topic. I’ll give me 2cents
Firstly- thanks to @chaoslabs for doing the hardest part of all by doing the initial proposal and research to kickstart conversation.
I agree with a lot of the counterpoints raised to the OP conclusions, including by @RealVovochka questioning the fairness of undisclosed formulas, and with @Callen_Wintermute underscoring the importance of Makers in the rewards.
- Why does Chaos Labs believe that the native V4 rewards program won’t be enough for this?
Lets clear this up first. The size of the rewards, while decent at $3.2mn a month, is actually low historically- and if volume takes time to ramp up as I expect it will, the estimate of $5.4mn in rebates paid will end up being a lot less. Current reward formula for example implies that if volume is very low and doesn’t move over, there won’t really even be much to give out. Obviously this requires kickstarting the flywheel.
I think $20mn is on the low end for what is indeed the largest need to incentivize user migration in the history of the protocol and is indeed under 5% of the treasury- less than 1% a month. The existing rewards program is simply giving back some rebates but that is not enough to kickstart a flywheel.
The move is much more significant and friction inducing than say a GMX v1 to v2 move. Everything not only has to move to a different chain, but having been the ones trying to help with the MM transition to v4 I can say first-hand that the liquidity providing tech is very different and requires a lot of dev time to implement properly. We go from standard limit orders to orders that expire every block so that they are more MEV-resistant and this already is totally new ball game.
I think $5-6mn a month for a pool that has to get divided between all participants would have been closer to whats needed to de-risk the transition of everyone including MMs, and $3.2mn will hopefully be enough.
These numbers do matter and are required- we have already seen since the most recent 50% reduction in LP rewards that we are up from 21 to 28% of rewards without changing anything(!)- showing that other MMs have switched off. In a world where they have to get their developers to figure out proper onchain trading we expect the MMs to quickly drop off further, and this would be a very real risk for the future of dYdX to lose that MM ecosystem diversity.
In summary-
It is not cheap for institutions to move to the v4 model, and though this might be abstracted for retail that just send market orders, the cost of decentralization and moving onchain as dYdX is envisioning does entail friction that needs to get compensated for in the first period to ensure survival through.
- Chaos Lab role
I am not worried so much about any nefarious dealings of chaos, even if Wintermute is an investor etc - everyone here has a reputation to protect. However-
- The formula being confidential
Here is where I would like to focus as this is the hardest topic. First off I fully agree that its far from ideal having rewards go out without transparency as to how that is done- indeed that is going against Defi’s general ethos. As a pragmatist however it is also clear to see that if it is public and gets gamed (as similar things have have happen before), that will just result in wash trading that isn’t good for anyone. Both of these points have been made clearly above by others.
As a compromise third option, we can try to create more transparency for the community, and also target more carefully the users and behaviors we need for a flourising v4. Here are some ideas-
a) Half of the monthly rewards going to passive side- to the Market Makers based on a ranking or for meeting some %market share threshold.
This would result in a lot of the usual end-of-epoch trading. One way to counteract is to have a private/randomized decision on what day/time out of say a 7-day window an epoch would reset. The behavior this would result in more natural from the Liquidity providers, and would just incentivize them for being a meaningful part of the market- creating a natural competition between them.
One way would be for the 50% of pool (currently $1.6mn of rewards) given out to the top 8 LPs in a set prize pool, and again with the hidden randomization on the exact end date-
1st place- 33%
2nd place- 24%
3rd place- 18%
4th place- 12%
5th place- 7%
6th place- 4%
7/8tht place- 2% each
The idea of adding other params like Open Interest, aum etc are as discussed over the years too gameable and just create an unfair advantage for those players with large balance sheets without them needing to quote competitively. The other param that is legitimate of showing market depth is generally great to have but in this case harder to do given the expire-by-block order types and the MEV considerations in v4.
Therefore a leaderboard for MMs similar to what Injective has used might be a good setup for passive side, and for taking side I would again just do volume based but in this case keep the same randomization on the monthly expiry so that it can’t get gamed at the end.
If there was sybil resistance we could do a square root on the volume before applying the pro-rata but with onchain wallets it would actually benefit sybillers and extractooors if we give more to smaller wallets than their pro rata share.
I agree with the wash trading monitoring and blacklisting wallets with clearly bad behavior.
I think then the only other item to decide for the formulas, especially on passive side is how to treat smaller markets outside of BTC/ETH. Historically it is important to give them separate consideration for rewards, otherwise if its just overall market % it drowns out the smaller markets and with no incentive to quote them they become illiquid and die. With completely new books needing to get built for passively quoting for functioning markets, PLUS with the added threat of MEV on v4, its important to have dedicated rewards for longer tail assets.
The question is do these get publicly shared or do the per asset % get kept hidden also? I think public is better for this item also!
In summary, the only item I would keep private is the epoch length- this can be randomized and kept private within some range of time and it would stop a good amount of the wash trading.
On the taking side, I think its ok to keep it more simple and base it on volume. I would only keep 30-40% of the extra $20million pool for taking, and certainly not more than 50%. Maybe 30% for taking pro rata on volume and ~10% on incentives targeting retail users and other programs (like these hedgie leaderboards).
As @CipherLabs has mentioned, reduction to rewards has led to decrease in liquidity and volumes, and its time now to make sure V4 starts on the right foot- Giving more transparency to the formula is more likely to bring users in as they will know what to expect and this is priority no.1. Avoiding wash trading is also a priority but since it can also get addressed through the monitoring and blacklisting, I think we should err on the side of transparency.
Thanks and happy to reply to others feedback to this.
