On July 24, ICON provided an update to the Yellow Paper, in a New Version 2.0 (IISS 2.0). This update came roughly 5 hours after ICON announced the highly anticipated date that staking on the ICON network begins: August 26. The IISS updates came as a bit of a surprise to many of us in the ICON community. I shared several conversations on Telegram with multiple ICONists and other P-Rep candidates. I’ll be up front in stating that my initial reaction to the update was negative, and matched the attitude of many in the community. This was mostly due to a misunderstanding of the details for a potential burn of 6% of staked tokens. After gaining further understanding of the updates and reasons for the decisions, however, my opinion has changed to positive. I think ICON made an excellent choice in the areas described in the paper and I support this paper and the policies set by ICON in it. My primary area of skepticism was based around a new Penalty System. Many other ICONists in the community shared my skepticism, as we had very lively discussions in various channels on Telegram. I provided an update to my conversion and why I changed my mind in those channels. The goal of this article to provide details for other ICONists who are skeptical of the Penalty System, describe why I changed my mind to support this proposal, and provide details so they can better understand and make an informed decision of where they stand.
There are many updates in the new Yellow Paper, but the one I want to focus on is the Penalty System. The medium article highlighting the Yellow Paper changes describes this system as such:
“ICON’s penalty system now has a token burning component. There are 3 types of penalties outlined in the IISS paper: Validation Penalty, Low Productivity Penalty, and Disqualification Penalty.
If a P-Rep suffers a Low Productivity Penalty or a Disqualification Penalty, 6% of the delegation toward such a P-Rep will be burned. This makes it extremely important for ICONists to delegate towards reliable and quality P-Reps.
We encourage ICONists to delegate across multiple P-Reps to diversify the risk of having their tokens burned.”
ICON is a Delegated Proof of Contribution system and requires all parties to contribute to be successful. I strongly believe there needs to be a penalty to incentivize good behavior of P-Reps and sound voting and incentives for iconists to vote for good P-Reps. Accountability is key in any ecosystem and penalties for missing the mark help enforce this. However, my initial thoughts were that this was a very harsh penalty to go after ICX of ICONists. I sought to further understand what situations would allow this.
The penalties that warrant a burn of 6% of the delegated ICX are:
- Low Productivity Penalty: “Applies when the Productivity Ratio of a specific representative has dropped below 85%. The Productivity Ratio is the ratio of actual blocks produced and validated divided by the number of opportunities to produce and validate a block.Representatives will not be subject to this penalty for their first 86,240 blocks as a block producing Representative”
- Disqualification Penalty: “Applies when a specific representative has been disqualified via a ‘P-Rep Disqualification Proposal’”
A disqualification proposal requires at least 66% of peer P-Reps to vote to disqualify a P-Rep. This is a serious proposal and is reserved for significant blunders. I think this is safe to assume that if you vote for good P-Reps who are producing fruit for ICON, then there is no need to worry about this type of penalty. The remainder of this paper focuses on the Low Productivity Penalty.
The Yellow Paper describes the Low Productivity Penalty as one in which the Productivity Ratio of a P-Rep drops below 85%. The paper does not describe the period of this Productivity Ratio. In light of this, there was much discussion on Telegram as to what the period of time is. If the period is short, then a P Rep could be producing very well for long stretches of time and then have an off day and suffer a Productivity Ratio. I thought this was extremely harsh and not fair to ICONists or P-Reps. I thought that a better option would be to eliminate future reward opportunity for the P-Rep and ICONists who delegate to that P-Rep for a period of a time (e.g. 2 months).
Scott Smiley of ICX Station was very helpful in chiming in and providing further clarity to my question of why this current method of penalty was chosen. He also listened to the community concerns and the alternative proposal. The most critical of points he clarified is that the Productivity Ratio is an aggregate of the entire time a team has been a block producing P-Rep. Thus, a team just needs to stay above 85% for their duration of block production, and this penalty will never happen. AWS compute service commitment provides a 99.99% uptime for their compute services. As long as the P-Rep properly configured their system, they should be well above the 85% requirement.
Additionally, if ICONists stake with multiple P-Reps, then the impact of one of these P-Reps falling under the 85% requirement will have less overall impact on the ICONists portfolio. Example:
- If I have 1000 ICX and stake to 1 P-Rep and that P-Rep has a Low Productivity Penalty, then I lost 60 ICX (6% of my initial ICX).
- If I have 1000 ICX and evenly stake to 10 P-Reps each (100 ICX each), then if 1 P-Rep has a Low Productivity Penalty I only lose 6 ICX (0.6% of my initial ICX).
It is advantageous to the ICON network to have a diverse set of P-Reps. We encourage voting for multiple P-Reps, as we discuss different qualities of P-Reps and give an overview in this article, also the art of P-Rep selection in another article.
While the penalties result in severe penalties via a 6% token burn, it is a rare occurrence for exceptionally poor behavior and it ensures that a malicious P-Rep does not just take their tokens and register again as a new P-Rep. Without burning tokens, this same offense could be repeated. There may be better ways to go about penalties, but I believe ICON thoroughly thought this one through and it is well architected. It better options are made known and agreed upon, then P-Reps can change the penalty system through Governance.
In conclusion, there needs to be incentives for P-Reps to perform their duties as node operators, ways to get rid of malicious of bad P-Reps, and also methods for ICONists to be incentivized to vote for good P-Reps. The new IISS system outlined in the Yellow Paper presents penalties to hold the system accountable as well as encourage diversification of ICX staking with multiple P-Reps. As long as ICONists are attentive and stake with multiple P-Reps, there should be very little risk of losing ICX. ICON is a Delegated Proof of Contribution network, and this penalty ensures ICONists are contributing and through valuable votes for good P-Reps. In return, the ICONists receive rewards, ranging from 6% to 36%, as is outlined in the Yellow Paper (we also give an overview of staking in this article).