KVORUM
compoundSource: Compound governor alpha · #35

COMP Speed Recalibration

executedJan 11, 2021, 05:11 PM UTCby660x6626…c35c

01Summary

AISummary by Kvorumclaude-haiku-4-5

Proposal to recalibrate COMP token distribution speeds across 9 active Compound markets based on 6 months of usage data. Each market receives a fixed 10% allocation plus variable rewards, enabling future A/B testing of incentive mechanisms. Implementation involves adjusting COMP emission rates across multiple market contracts.

Type
parameter change

Key changes

  • Recalibrate COMP distribution speeds across markets based on 6-month usage data instead of arbitrary initial parameters(high)
  • Implement fixed 10% allocation of COMP rewards spread evenly across 9 active markets(high)
  • Split market rewards into fixed and variable quantities to enable A/B testing of optimal incentive levels(medium)
  • Adjust individual market COMP emission rates ranging from 1.95e15 to 67e15 wei per block(medium)

Notable concerns

  • Initial COMP distribution parameters were set arbitrarily and required reset
  • Proposal relies on community forum discussion for technical details and rationale
Contracts0x3d9819210a31b4961b30ef54be2aed79b9c9cd3b0x4Ddc2D193948926D02f9B1fE9e1daa0718270ED50xC11b1268C1A384e55C48c2391d8d480264A3A7F40x39AA39c021dfbaE8faC545936693aC917d5E75630xf650C3d88D12dB855b8bf7D11Be6C55A4e07dCC90x5d3a536E4D6DbD6114cc1Ead35777bAB948E36430x6C8c6b02E7b2BE14d4fA6022Dfd6d75921D90E4E0xB3319f5D18Bc0D84dD1b4825Dcde5d5f7266d4070x35A18000230DA775CAc24873d00Ff85BccdeD5500x70e36f6BF80a52b3B46b3aF8e106CC0ed743E8e4
Generated by Kvorum
model: claude-haiku-4-5
provider: anthropic
profile: anthropic-fast-v1
prompt: v1.0
generated: 2mo ago
Read the description ↗

02Description

241 words

COMP Speed Recalibration

Summary

As Robert Leshner mentioned in his forum post, the community now has the power to control COMP distribution across markets after Gauntlet's recent proposal (#33). This will allow the protocol to better and more efficiently manage how COMP is used to improve and further the protocol's interests. Given that the initial COMP distribution numbers were set somewhat arbitrarily, it has become clear that it is worth resetting the parameters based on usage over the last six months. The refreshed numbers proposed will be used for A/B testing optimal COMP emission quantities in future proposals.

In a comp.xyz forum post, Robert and Sirokko have put forth reasonable suggestions that we believe would improve upon the existing values. Both of their values for COMP distribution were quite similar and we chose Sirokko's numbers for two reasons:

  1. All markets are given a fixed quantity (10% of rewards spread evenly across the 9 active markets) of COMP rewards
  2. Splitting each market's COMP rewards into a fixed quantity (common to all markets) and a variable quantity makes it easy to analyze improvements in future A/B tests for incentives

Please read the forum post for more details.

Acknowledgments

  • Robert Leshner: For starting the discussion and putting out the first distribution numbers
  • Sirokko: For providing a coherent methodology for this refresh
  • Hsien-Tang Kao: For analyzing 1-step recursive borrowing in existing markets

References

03Mismatch analysis

AIMismatch analysis by Kvorummedium confidenceclaude-sonnet-5
Consistent

The description explains that this proposal recalibrates COMP distribution speeds across the 9 active Compound markets, using a fixed+variable methodology proposed by Sirokko, and explicitly defers exact numbers to a linked forum post rather than stating them in the proposal body. The calldata consists of exactly 9 _setCompSpeed calls, one per market address, each setting a new COMP emission rate — this structurally matches the described action (a market-wide speed reset) in both scope (9 markets) and mechanism (setting speeds via the standard Comptroller function). There is no target mismatch, no unrelated action, and no evidence of value manipulation beyond what the description authorizes, since the description intentionally does not enumerate specific values and instead points to the forum post as the source of truth. Because the specific COMP speed values are not stated in the prose itself, a full numeric cross-check against 'Sirokko's numbers' is not possible from the provided materials, which slightly lowers confidence but does not constitute a discrepancy per the guidelines (summarization deferring detail to a linked source is legitimate). Overall, the on-chain actions are consistent with what the proposal claims it will do: reset COMP emission speeds across the nine active markets per the community-discussed methodology.

Discrepancies

  • lowomitted in description

    The proposal text does not restate the actual numeric COMP speeds being set (it defers entirely to an external forum post). This is a legitimate summarization for a recalibration proposal, but it means the specific values (e.g. 0.067, 0.01075, 0.00195, 0.005 COMP/block) cannot be verified word-for-word against the prose itself — only against the referenced forum post, which is outside the scope of the on-chain description provided.

    “Please read the forum post for more details.”

Generated by Kvorum
model: claude-sonnet-5
provider: anthropic
profile: anthropic-strong-v1
prompt: v1.0
generated: 2mo ago
Review the decoded actions ↗

04Actions

9 total
#0_setCompSpeed(address,uint256)chain 0x1
Target
0x3d98…cd3b
Value
0 wei
Function
_setCompSpeed(address,uint256)
Arguments
{
  "0": "0x4Ddc2D193948926D02f9B1fE9e1daa0718270ED5",
  "1": "10750000000000000"
}
Calldata
#1_setCompSpeed(address,uint256)chain 0x1
Target
0x3d98…cd3b
Value
0 wei
Function
_setCompSpeed(address,uint256)
Arguments
{
  "0": "0xC11b1268C1A384e55C48c2391d8d480264A3A7F4",
  "1": "10750000000000000"
}
Calldata
#2_setCompSpeed(address,uint256)chain 0x1
Target
0x3d98…cd3b
Value
0 wei
Function
_setCompSpeed(address,uint256)
Arguments
{
  "0": "0x39AA39c021dfbaE8faC545936693aC917d5E7563",
  "1": "67000000000000000"
}
Calldata
#3_setCompSpeed(address,uint256)chain 0x1
Target
0x3d98…cd3b
Value
0 wei
Function
_setCompSpeed(address,uint256)
Arguments
{
  "0": "0xf650C3d88D12dB855b8bf7D11Be6C55A4e07dCC9",
  "1": "9650000000000000"
}
Calldata
#4_setCompSpeed(address,uint256)chain 0x1
Target
0x3d98…cd3b
Value
0 wei
Function
_setCompSpeed(address,uint256)
Arguments
{
  "0": "0x5d3a536E4D6DbD6114cc1Ead35777bAB948E3643",
  "1": "67000000000000000"
}
Calldata
#5_setCompSpeed(address,uint256)chain 0x1
Target
0x3d98…cd3b
Value
0 wei
Function
_setCompSpeed(address,uint256)
Arguments
{
  "0": "0x6C8c6b02E7b2BE14d4fA6022Dfd6d75921D90E4E",
  "1": "1950000000000000"
}
Calldata
#6_setCompSpeed(address,uint256)chain 0x1
Target
0x3d98…cd3b
Value
0 wei
Function
_setCompSpeed(address,uint256)
Arguments
{
  "0": "0xB3319f5D18Bc0D84dD1b4825Dcde5d5f7266d407",
  "1": "1950000000000000"
}
Calldata
#7_setCompSpeed(address,uint256)chain 0x1
Target
0x3d98…cd3b
Value
0 wei
Function
_setCompSpeed(address,uint256)
Arguments
{
  "0": "0x35A18000230DA775CAc24873d00Ff85BccdeD550",
  "1": "1950000000000000"
}
Calldata
#8_setCompSpeed(address,uint256)chain 0x1
Target
0x3d98…cd3b
Value
0 wei
Function
_setCompSpeed(address,uint256)
Arguments
{
  "0": "0x70e36f6BF80a52b3B46b3aF8e106CC0ed743E8e4",
  "1": "5000000000000000"
}
Calldata

05Tally

Summed from votes

Voters0
VP participating0
Outcome (current)—
Current support—

06Voters

0 total
Filter:
VoterChoice% of totalRationale
660x6626…c35cFor126.1K0.00%—6y ago
AC0xac57…faf6For105.1K0.00%—6y ago
540x54a3…e189For100.1K0.00%—6y ago
7E0x7e4a…19faFor100K0.00%—6y ago
450x458c…730eFor50K0.00%—6y ago
E50xe5b5…4b03For29.9K0.00%—6y ago
870x879d…0f11For22K0.00%—6y ago
DAdakeshi.eth0xe3c9…6b21For20K0.00%—6y ago
140x1494…3c2bFor18.8K0.00%—6y ago
7F0x7f89…58c4For12K0.00%—6y ago
5B0x5b97…66c2For10K0.00%—6y ago
B00xb032…449bFor9.1K0.00%—6y ago
580x582c…d468For9K0.00%—6y ago
8D0x8d07…a265For7.5K0.00%—6y ago
B80xb8a6…6406For5K0.00%—6y ago
900x905c…e25aFor5K0.00%—6y ago
FC0xfc04…5ac2For2.3K0.00%—6y ago
7F0x7f84…4d0bAgainst158.20.00%—6y ago
9B0x9b68…2984Against100.10.00%—6y ago
030x03f0…7618Against20.30.00%—6y ago
840x84e3…5a95For10.80.00%—6y ago
E00xe0b3…2514For80.00%—6y ago
930x9348…0c53For2.20.00%—6y ago
FB0xfb30…ea3fFor1.30.00%—6y ago
420x42ed…d691For10.00%—6y ago
850x8579…213aFor0.50.00%—6y ago
850x8590…1637For0.30.00%—6y ago
020x02af…6516For0.30.00%—6y ago
JTjthrack.eth0x44c6…aabdFor0.30.00%—6y ago
F60xf67e…7bf6For0.20.00%—6y ago
3E0x3e61…322bFor0.20.00%—6y ago
A00xa0e4…ea1eFor0.10.00%—6y ago
400x4007…7f4eFor0.10.00%—6y ago
6F0x6f7a…c97bFor0.10.00%—6y ago
2E0x2eb4…8c26For0.10.00%—6y ago
000x0006…13f5For00.00%—6y ago
A40xa455…eff3For00.00%—6y ago
E00xe02c…5e3eFor00.00%—6y ago
1F0x1f76…f6ddFor00.00%—6y ago
H0h0neybad9er.eth0x26a2…4cbeFor00.00%—6y ago
Showing 40

07Forum

AIForum synthesis by Kvorumclaude-sonnet-5
sentiment · mixedthread · constructive

Arguments for

  • The current COMP speed allocation is an arbitrary relic of when Proposal 033 passed and should not be anchored to as a baseline; governance now has responsibility to actively (re)calibrate it.— rleshner, haseebq
  • COMP emissions should be optimized using rigorous, data-driven experiments (A/B-style hypothesis testing) rather than left static, since $125M/year in incentives is a large spend whose ROI is unclear.— tarun, haseebq, jthrack
  • Distribution should shift away from 'farmers' who exploit recursive leverage/high-rate markets toward genuine long-term users and governance participants.— rleshner, Sirokko, dabar90, jthrack
  • Important collateral assets like ETH and WBTC receive negligible COMP relative to their importance and should get a meaningful baseline allocation to change user behavior and deepen collateral markets.— rleshner, dabar90, Sirokko
  • Increasing the COMP allocation to the COMP market itself acts like a staking reward that incentivizes long-term holding rather than immediate selling.— rleshner, tarun, Sirokko
  • COMP distribution's core purpose is to decentralize protocol ownership, not merely minimize customer acquisition cost, so reallocation should be judged by its effect on decentralization.— aklamun, jthrack
  • Similar assets (e.g., BAT/ZRX/UNI or stablecoins) can be grouped and standardized to simplify and rationalize the speed-setting process.— rleshner, dabar90, Sirokko

Arguments against

  • The existing distribution 'isn't broken' and intuitively tracks interest rates, so major changes should wait until effects of other initiatives like vesting are observed.— lay2000lbs
  • Manually adjusting COMP speeds via governance controlled by a small number of large token holders undermines decentralization and looks like 'big business' behavior rather than systematic, transparent adjustment.— Fishbtc, massnomis
  • Reducing rewards to recursive-leverage markets like DAI may simply push farmers to recursively leverage other collateral markets (ETH, WBTC), artificially inflating borrow rates there instead of solving the underlying problem.— trader-geo
  • The proposal process was rushed (posted with a short voting window) and lacked the depth of academic/economic rigor warranted for a decision of this magnitude, and should be postponed for a more dynamic, decentralized economic model.— massnomis
  • The vesting/lockup mechanism that is supposed to complement the new speeds was not yet well-defined before the reallocation proposal moved forward.— massnomis, jthrack

Unresolved concerns

  • Whether reducing rewards to stablecoin recursive-leverage farming actually curbs farming overall or just displaces it into other collateral markets, inflating their borrow rates.— trader-geo, jthrack, aklamun
  • Governance allocation decisions are effectively made by a small number of large holders, raising decentralization/legitimacy concerns about the process itself.— Fishbtc
  • Whether protocol reserves and a treasury (funded by reduced COMP emissions or direct COMP sales) should be prioritized alongside or instead of speed recalibration was raised but explicitly deferred to a separate discussion.— Sirokko, jthrack
  • Design of a vesting or 'anti-farmer' penalty/staking mechanism to complement speed changes remained conceptual and unresolved.— mrhen, massnomis, jthrack
  • The inability to set different supplier/borrower split ratios (e.g., 75/25 for collateral markets) without a protocol code change was acknowledged as a limitation but left as a future improvement rather than resolved.— Sirokko, rleshner

Notable participants

  • rleshnerOpened the thread with guiding questions, synthesized community feedback into a baseline allocation proposal, and closed the thread noting the changes were enacted via Proposal 035.
  • tarunGauntlet researcher who proposed a rigorous experimental/hypothesis-testing framework and provided data analysis on recursive borrowing to inform allocation choices.
  • SirokkoProvided a detailed alternative allocation model with a fixed baseline per market plus targeted incentives, and argued extensively for treasury/reserve considerations.
  • jthrackAdvocated for net borrow/supply rate balancing and reallocating COMP toward development or treasury diversification to reduce farming incentives.
  • aklamunReframed the debate around decentralization of ownership as the primary goal and proposed a max(lend reward, borrow reward) mechanism to curb recursive leverage.
  • haseebqStrongly endorsed moving away from the arbitrary status quo toward an optimized, data-informed emission rate.
  • lay2000lbsRepresented a cautious, status-quo-leaning viewpoint urging minimal changes.
  • FishbtcRaised concerns about governance centralization and the optics of manual reallocation by a small group of holders.
  • massnomisCriticized the pace and rigor of the governance process and called for a more dynamic, decentralized economic model before passing changes.
  • dabar90Highlighted opportunity costs for ETH holders and questioned Compound's user growth relative to competitors, supporting reduced stablecoin incentives.
Generated by Kvorum
model: claude-sonnet-5
provider: anthropic
profile: anthropic-strong-v1
prompt: v1.0
generated: 2mo ago
Open the thread ↗