Anders Elowsson pfp

Anders Elowsson

@anderselowsson

78 Following
744 Followers
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
The appendix provides simple approximations of the validator count under a Zipfian distribution of stakers. The log adjusted D/(20*ln(D)) gets very close and the loglog adjusted measure is almost perfect, as indicated in the plot. Check out the post for further details!
0 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
The scale parameter f_y determines how much the yield should change given a specific consolidation force (the scale of the incentive). The appropriate scale depends on the currently offered yield and MEV, the perceived risk of being active as a staker, deposit size, etc.
1 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
Validators might congregate at sizes that produce the best risk-adjusted rewards and the protocol’s priorities might change with consolidation level. For these reasons, validators can as an alternative be given an f2 based on their position in a sorted list of validator sizes.
1 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
The blue line is equidistant in the log domain and the yellow line weighs together the green and red line, in this example by averaging them. These options are less “lopsided” and will see a more or less equal impact on f2 when a validator doubles in size or activity rate.
1 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
As an example, 4096 ETH can be divided into (A) 1 2048-ETH validator and 64 32-ETH validators or (B) 4 1024-ETH validators. With Orbit thresholded at 2048, active stake is roughly equal (so green sets A=B), but the protocol might prefer 4 validators over 65 (so red favors B).
1 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
The force distribution f2 is unique to each validator and varies with activity rate or size or both. It distributes the collective f1 across the validator set through f1*f2 to produce individual incentives. Several options are reviewed in the post, a summary plot shown here.
1 reply
0 recast
1 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
The green line estimates the assumed risks based on activity rate and is neutral w.r.t. the active stake of some endowment. The red line simulates a validator fee and is inversely proportional to validator size. Given some endowment, it is neutral w.r.t. occupied validator spots.
1 reply
0 recast
1 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
It consists of the two shape variables f1 and f2 in the range 0-1 and the scale variable f_y. The combined incentive attenuates the yield by f1*f_y(c+f2). The consolidation force f1 is the same for all validators and varies with consolidation level, i.e., validator count V.
1 reply
0 recast
1 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
In Orbit SSF, validators rotate based on size, leading larger validators to take on greater risks. We must therefore provide some incentives to ensure consolidation in the staking set. This post offers a framework for designing such incentives. https://ethresear.ch/t/consolidation-incentives-in-orbit-vorbit-ssf/21593
1 reply
3 recasts
24 reactions
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
Here is the change in issuance yield. Whether this automated graduated approach is worth the additional implementation complexity will depend on the shape of the new reward curve, the quantity of stake at the time of the switch, and the sensitivity among stakers to abrupt changes in yield.
0 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
This of course also means that there is a weaker guarantee that the quantity of stake is maintained close to current levels. The change in staking yield with 300k ETH of MEV would look as follows.
1 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
Another example using the purple reward curve from the Practical endgame on issuance policy-post is here shown, but the curve is adjusted to peak at an issuance rate of i=2^-8=0.39%. A lower peak gives stronger guarantees on the maximum amount of ETH that can be issued each year.
1 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
Issuance yield is instead affected as shown in this plot. The starting point and end point of the automated reduction can of course be adjusted.
1 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
The easiest way to do this, shown in these plots, is to slowly bring down the constant k in the equation for the new reward curve. When k is infinite, both curves are identical. Staking yield with 300k ETH of MEV per year would be affected as shown here.
1 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
If the quantity of stake continues to rise and we reduce issuance, the yield differential between the new and the current reward curve may become very large. We could then consider implementing an automated, gradual reduction down to a new reward curve over a period of 1-2 years.
1 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
My talk from Devcon on a Practical endgame on issuance policy is now up. I review the motivations, impacts and potential downsides of a reduction in issuance. This thread will visualize an automated gradual reduction in issuance, which I suggested in an answer at the end. https://www.youtube.com/watch?v=m91Wu6-cdwk
2 replies
3 recasts
34 reactions
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
Here is rough draft of how we can design consolidation incentives in Orbit SSF. Publishing it now since I will briefly refer to it during my talk at Devcon tomorrow. https://notes.ethereum.org/@anderselowsson/Incentives_SSF
0 reply
0 recast
5 reactions
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
Whereas a cap on the circulating supply is an intangible monetary policy, a cap on the issuance rate is not. We get the same memetic simplicity and possibly a deflationary native asset, without compromising security. The post concludes with a set of questions for the community.
0 reply
0 recast
0 reaction
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
The post also introduces a tangible framework for Ethereum that is easy to understand: never exceed an issuance rate of 0.5%. A stringent cap on issuance is desirable because it caps the inflation rate, solidifying ETH as trustless sound money with preserved economic security.
1 reply
3 recasts
2 reactions
Anders Elowsson pfp
Anders Elowsson
@anderselowsson
Practical endgame on issuance policy This post presents a practical endgame on issuance policy that can stop the growth in stake while guaranteeing proper consensus incentives and providing positive regular rewards to diligent small solo stakers. https://ethresear.ch/t/practical-endgame-on-issuance-policy/20747
2 replies
8 recasts
62 reactions