The case for term limits

English General Discussions
1

Staking is not lending out your Ada for interest payment. Staking is no different to voting.

Every time a pool produces a block all the stakers of that pool are collectively approving that block. In essence they are telling the rest of the Cardano network: We sign this block with the pool keys and this decision is supported by the weight of our collective stake.

Cardano security is built upon this assumption. It is a pivotal assumption that protects Cardano against double spending. The assumption is that so long as greater than 50% of participants, by stake weighting, act honestly then double spending can’t happen. Newer technologies like Mithril have further extended the dependence upon this critical assumption. Mithril permits faster block confirmation since if greater than 50% of participants by stake weighting agree a block, the protocol affirms that this block cannot get forked out.

But this crucial assumption starts to unravel a bit when the stakers are not actually approving the blocks.

Obviously, it is unreasonable to expect all stakers of a pool to be awake and online every second of the day in order to approve every block their pool produces. Thus Cardano has automated the mechanism to produce and sign blocks. Only the pool operator checks the blocks by running the protocol, and all the stakers have vouched with their staking keys that they trust this operator will faithfully perform the protocol. Nevertheless the assumption remains that the stakers of a pool continue to approve of that pool operator’s performance every time a block is made.

This is exactly like voting for a political representative. Unless you are a political animal, you are probably not going to want to know about every decision your representative takes, nor hear every negotiation, nor read every argument about every topic. You haven’t got time, you have your own life issues to deal with. This is why you vote to appoint a political representative who you believe will act to protect your interests. But we all know that people change, and their views change, over time. So you might want to change your appointed representative next election. But see; that is the key. Every democratic system has term limits for exactly this reason. No democratic system thinks it is a good idea to have participants lock in support for a representative in an unbounded manner, to automate away the voting process… forever. Every democratic system has term limits and participants are required to actively re-affirm their chosen representative, by voting.

Staking is no different. We need term limits.

These are some questions I have been pondering while trying to understand how much “sticky stake” might compromise Cardano’s consensus:

  • What percentage of new users select a stake pool without knowing anything about the pool just from the front page of their wallet? What percentage simply follow a youtube recommendation? 10%, 20%, 30%.
  • How will that percentage change over time as the staking yield reduces towards 2% or less.
    Won’t care factor reduce as rewards reduce?
  • What sort of pools are likely to capture the most “sticky stake”? IE: Capture the stake that does’t care what happens and won’t re-stake, or the stake that is more likely to lose their keys through inattention to detail / general disinterest.
    Is it the youtuber with the big mouth pumping meme coins, or the youtuber talking about technical design features deep in the weeds? What about the pool that doesn’t use youtube advertising?
  • What happens if the pool operator decides to sell his pool keys to someone else because he sees some other chain has better memes?
  • How much are his pool keys worth if he has more “sticky stake” that won’t move when fees are increased?
  • If a multi-pool operator has lots of “sticky stake” doesn’t this give him more options to split his pools and charge different fees on different pools? Doesn’t the “sticky stake” give this pool operator a competitive advantage that was not envisaged by the protocol designers?
  • If a pool operator wanted to optimise for accumulating “sticky stake” how might he go about this?
    Maybe start off with one pool and a gimmicky youtube channel presented by a celebrity to attract the hype crowd. Then after a while you spit the pool and charge a higher fee on the original pool and much lower on the new pool. Rinse and repeat. Each time you split and charge a higher fee on the original pool you hive off the “sticky stake” in that pool, but you continue to accumulate new entrants in the new pool with lower fees. You can afford the lower fee on your new pool because you are profiting from the “sticky stake”, plus you get economies of scale from multiple pools using the same infrastructure.

I think we should consider term limits on staking, which is really consensus vote delegation, just like there should be term limits on governance vote delegation. IE: When you stake you sign that you support a stake pool for a finite period of time. After this period of time your approval for this pool as your representative for blockchain consensus expires and you need to re-stake or re-affirm your pool selection.

And what’s even better is that having term limits can actually improve the staking yield for active participants. If we no longer reward the sticky stake that does nothing, and actually contributes negatively by making Cardano less “anti-fragile”, then there can be more rewards allocated to the active participants.

2 Likes
2

Thank you for bringing this to the attention of the community. I believe your concerns are valid & I totally agree with your proposal on limit terms on staking. I’d love to see other people’s take on this matter…

2 Likes
3

I would like to see this potential network feature incorporated into any long term plan to deal with these consequences of “sticky stake” as a reliable means of notifying delegators, pool operators, and perhaps analytic services (click rendered proposal in first comment to read the document):

2 Likes
4

I agree with you

2 Likes
5
  • What percentage of new users select a stake pool without knowing anything about the pool just from the front page of their wallet? What percentage simply follow a youtube recommendation? 10%, 20%, 30%.

I suspect quite alot, much more than I would have ever hoped for after researching Cardano and becoming an ADA holder the first weeks Bittrex & Binance provided me access to ADA.

  • How will that percentage change over time as the staking yield reduces towards 2% or less.
    Won’t care factor reduce as rewards reduce?

I didn’t think so initially. I do start to get this feeling now however that care factor may drop off.

  • What sort of pools are likely to capture the most “sticky stake”? IE: Capture the stake that does’t care what happens and won’t re-stake, or the stake that is more likely to lose their keys through inattention to detail / general disinterest.
    Is it the youtuber with the big mouth pumping meme coins, or the youtuber talking about technical design features deep in the weeds? What about the pool that doesn’t use youtube advertising?

Originally I looked for SPOs who had a highly technical background, may not have been well known, but were technologically savvy and/or had extensive Linux systems engineering experience, Site Reliability and monitoring/alerting skills, and experience in both cloud and datacenter infrastructure. That said, I come from IT and technical background.

In the end I think the answer to your question, many years after Shelley and stake pools arrived, tends to be either the youtuber with a big mouth, or a combination of the youtuber with a big mouth with one who talks a little bit about technical design. My personal view, and some discussions at events with average delegators, those who get sticky stake are actually probably much less technically savvy than they portray themselves as, but they talk a good game and that is all it takes for the average ADA holder who is extremely non technical to be influenced and sticky.

The rest of your questions are great. I’d love to hear responses from others on these topics. I have some opinions on them, but don’t want to monopolize this discussion.

Overall, I think the game theory concepts suggested that delegators would be more involved directly and pay close attention to the rewards, moving stake around as needed, preventing the “sticky stake”.

After years of delegating, research, and reading Twitter / watching YouTube, today I feel that:

  1. The game theory was way overblown as to how much involvement the average ADA holder would have, compared to the reality we see today.
  2. The 2 epoch delays between block creation and rewards impacted the average ADA holder enough to detract from the game theory research that was presumed to keep delegators moving stake around.
  3. The SPOs being pushed to to market themselves as businesses via Twitter and YouTube, essentially becoming or teaming up with influencers, added to this “sticky stake” model.

I think the game theory had a small chance of proving itself true, but so many other things were tossed into the mix that the concepts of the game theory on delegation fell quite short of what the theoretical research suggested it would achieve.

I’m not against your suggestion of term limits, if the difference actually improves the staking yield for active participants. However, if the protocol keeps active participants around 2% (once it gets that low) while the difference is slurped into the treasury, continuously depleting the reserves at the same level, then I would be against this.

Somewhat Off Topic:
I fully support the existing deflationary model. I also support funds going into the treasury for development and to cover costs, but in no way do I support increasing the amounts going into the treasury beyond the amounts it already gets, I think it could have easily been less for the last 4 years and the APR/ROS remaining higher for longer would have been a big influencing factor for to attract new wallets/ADA holders.

Poor performance of pools, lack of holders participating in delegation (of staking or governance) after term limits expire, etc. I do NOT feel should increase the treasury share of the rewards per epoch. Either an increase goes to active participants, or the inactive participants amounts should stay in the reserves to extend the staking rewards for active participants.

In fact I’d love to see some DRep out there suggest a reduction to the Treasury % of epoch rewards, nothing drastic, just a reasonable adjustment that either takes the leftover % and splits it among active delegators, or returns it to the reserves.

2 Likes