[Signal Request] Should We Introduce a Performance Modifier to the Delegate Compensation Trial?

Overview

The GovAlpha Core Unit would like to propose a change to the Delegate Compensation Trial that is currently taking place.

Please see MIP61 and these two previous Signal Requests for further information:

We would like to propose a change to the cut-off value for delegate compensation. At present, a Recognised Delegate who does not have both a participation metric and a communication metric above 90% will not be eligible for compensation. Up-to-date metrics for each Recognised Delegate can be viewed on the voting portal.

We believe that this cut-off is potentially more extreme than necessary. Therefore, the following proposal allows for the Recognised Delegates to have more flexibility whilst still incentivising Recognised Delegates to push to achieve 90% in both metrics.

Proposed Changes

We are proposing to introduce a Performance Modifier which will result in a reduced amount of compensation at lower levels of metrics with the following changes:

  • Delegate compensation will continue to be calculated per the formula detailed in MIP61, with a Performance Modifier applied to the final amount, based on the Recognised Delegates lowest scoring metric.
  • Recognised Delegates with both metrics above 75% will become eligible for compensation.
  • If either metric is below 75%, the Recognised Delegate will receive no compensation.
  • There will be a Performance Modifier defined by a linear scale starting at a 40% Performance Modifier applied at metrics of 75%, rising to 100% Performance Modifier for metrics of 90% and above.

This is demonstrated in the below graph:

Example

To demonstrate the effects of this new Performance Modifier, we have included the following example. These numbers are not based on the performance of any Recognised Delegate but are purely illustrative:

Delegate Participation % Communication % MKR Delegated Performance Modifier Monthly Compensation
Delegate A 95 95 12000 100.00% 12000.00
Delegate B 95 85 12000 80.00% 9600.00
Delegate C 95 95 8000 100.00% 10733.13
Delegate D 70 80 8000 0.00% 0.00
  • Delegate A receives maximum compensation of 12000 DAI as they have more than 10,000 MKR delegated, and both metrics are above 90%.
  • Delegate B receives 9600 DAI as they have a Performance Modifier of 80% applied to their compensation because their lowest scoring metric is 85%.
  • Delegate C receives 10733.13 DAI. This is less than Delegate A because they have less than 10,000 MKR delegated, but more than Delegate B because Delegate B’s Performance Modifier is 80%.
  • Delegate D receives 0 DAI. This is because their lowest scoring metric is 70%, so the Performance Modifier is 0%.

We believe that this Performance Modifier increases the fairness of compensation for Recognised Delegates while still encouraging them to have metrics as high as possible. The above example shows the importance of higher metric scores by awarding Delegate C more compensation than Delegate B despite having 4000 fewer MKR delegated.

Poll

Should We Introduce a Performance Modifier to the Delegate Compensation Trial?
  • Yes
  • No
  • Abstain

0 voters

Next Steps

This poll will close on Friday, December 10 at 20:30 UTC+0. If a majority of forum voters support this change, this will move forward to on-chain polling the following week.

Notes

  • A positive outcome for this Signal Request does not necessarily result in an update being applied to MIP61 but only impacts the ongoing Delegate Compensation Trial. We hope that the author of the MIP will recognise the view of Maker Governance and include these changes in the MIP should this Signal Request be successful.
  • This change will only be applied prospectively. Therefore, it will not retroactively apply to the Delegate Compensation calculations for November.

cc: @Recognized-Delegates @GovAlpha-Core-Unit @psychonaut

8 Likes

I also think this is a sensible way to go. The danger is that delegates dropping below 90% can fall into a negative feedback loop where it is not worth their time / gas to vote and communicate, because their stats are low enough to be hard to recover from.

Delegates can be especially prone to this early on, as we’ve seen with a couple of the delegates. It’s easier to make mistakes when you’re new, and we don’t want delegates to accidently miss a few votes, then miss out on compensation for potentially multiple months.

The proposed scale mitigates this issue by more gradually applying the penalty as metrics decrease.

It also makes involuntary absence easier to deal with and recover from. And frankly it also allows delegates to take a week off and not worry about being too harshly penalized.

The flip side of this is that it makes the compensation calculation slightly more complicated by adding an additional term. Given that we can automate calculation and share links to the google sheet calculating compensation (including comments to explain each stage of the calculation) this feels like a worthwhile trade-off right now, especially when we’re in an ‘attract delegates’ phase.

2 Likes

Rest assured :slight_smile:

1 Like

Off topic, but I wish delegates could attach their communication to the votes instead of putting them in the forums. Communications in the forums are organized by delegate, but what I really want to see are justifications organized per vote. If there is a controversial vote then I want to see how the delegates voted and why.

6 Likes

Theoretically, you can submit a comment when you vote and then we could organise that by vote in a UI somewhere, but then that introduces extra complexity for the delegates

How so? If the justification forum post is no longer required then the delegates are still only reporting their justification once.

Are communication and voting metrics an “all-time” score or are they reset regularly e.g. every month? If it is the latter, then I don’t think delegates fall into negative feedback loops or have initial mistakes during setup affect them in the long-term.

My view on this is that I think it is quite reasonable to expect delegates to vote and communicate at a high rate like 90%. Given that the compensation increases steeply from 75% to 90%, the financial incentive to be above 90% is very similar in both the current system and the proposed system.

While it might make sense for the trial period, I doubt this change will affect delegate behavior much.

They are all-time and not reset. A month doesn’t give a reasonable period for MKR holders to get a good picture of how much a delegate will participation.

We could theoretically reset them annually, or over a different longer period, but I think this makes the most sense for now.

1 Like

Oh! I had assumed you were doing a three or six month moving average or similar. Using an all-time average could lead to bad outcomes if a delegate performs well for a year and then underperforms. It would take a long time for the stats to reflect recent performance.

Yes true, but if they drop off the face of the earth without giving us a heads up, we’ll probably just delist them as recognized delegates and stop paying them.

If they just get a bit spotty, you’re right, it could be bad. Maybe something like min(all-time, quarterly) or something makes sense in the long run. Or just a moving average as you suggested.

1 Like

+1 to a moving average that’s on a reasonable time frame. Whatever is easiest on gov-alpha has my vote though.

1 Like

This signal closes tomorrow and ideally we would see a bit more participation. If you haven’t submitted your opinion, please consider doing so.

Whilst still getting my sea legs, I’m strongly in favor of a scheme that shares compensation with part-time delegates or those who reach a reasonably high bar of participation with added incentives to exceed the minimum expectations. Directionally this proposal looks like a good adjustment to the status quo.

Best regards,
Kianga

1 Like

@GFXlabs and @MakerMan, I don’t want to derail your delegate platform threads but I just wanted to point out that there is no change to compensation for delegates whose metrics are between 90 and 100%, as a result of this SR.

I don’t want you to think you will be adversely effected by voting for this modifier. The only effect of this SR is allowing delegates with metrics between 75 and 90% a reduced rate of compensation.

As per the text of the poll:

when the lowest scoring metric is 90% or greater, compensation will be 100% of the eligible amount.

1 Like

See MIP61: Delegate Compensation - #27 by psychonaut

2 Likes

This Signal Request passed on-chain and the Performance Modifier will be used by GovAlpha when calculating Delegate Compensation for the remainder of the Delegate Compensation Trial.

The on-chain vote can be seen here.

Thank you to everyone who participated in this Signal Request.

2 Likes

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.