RE: Gridcoin Research 4.0-2018 Roadmap Progress Report
I have been thinking about CBR. I am nervous that CBR may actually end up being a disincentive for large balance holders and could have adverse economic implications. (Right now the current reward PoS reward mechanism encourages large balances but does not encourage keeping the wallet online. We have to be careful not to exchange one problem for a worse one economically, which would be to unwittingly create a disencentive for large balance holders to keep large balances of gridcoin.)
I think a good compromise would be a PoS reward that is linearly related to the average gridcoin balance at the time of staking, but NOT a function of the time since last staking (other than the indirect impact of staking frequency). This appropriately rewards large balance holders, but also encourages everyone to keep their wallets online for staking, since staking more often with this method will lead to greater rewards. Keeping the wallet offline will prevent staking and result in no reward during the offline period. Since the staking weight would still be essentially proportional to the balance compared to the total gridcoin in circulation, the staking frequency (and reward) would still be proportional to the average balance maintained, which would further encourage whales to stay online which puts forth large staking reserves.
Could you explain how you see CBR as a possible disincentive? Also, what are the adverse economic implications that make you nervous?
I am a bit confused to the solution you propose as well. Could you rephrase or give an example?
See my reply below, my idea of scaling both the reward and frequency results in essentially a quadratic reward curve. That may tilt things too much towards large balance holders. What I think by disencentives is for whales is that we have to understand WHY they are holding large balances. Is it the interest? I am not sure the whales really would have such a problem keeping the clients online to stake as much as their concern about the amount of interest they are collecting. The CBR needs to be carefully calibrated to provide the equivalent of the same interest they are getting now, but with the requirement that they keep their clients online to get it. The CBR value that has been discussed seems an arbitrary number though. Maybe it was picked through a calibration analysis like that and it just came out at that even number. The frequency calculation based on relative magnitude seems pretty set, so the only way to calibrate the CBR would be to create a scale factor which adjusts CBR periodically to match the desired system-wide stake payout with the previous total interest payments. This should cause participants to get almost the same payment as the interest they did before, as long as they keep their clients online.
I like to hope that investors invest because of the blockchain. If they invest for interest without faith in the blockchain, they would be investing in something for more of that thing which they see 0 value in.
There have been no CBR values proposed. Those in the article are place holders.
I agree, that we need to calibrate CBR value based on our intent -- what that intent is still needs to be defined.
The frequency to stake is currently and for the immediately future based on balance, not magnitude.
Low balance users not staking is a problem across the board with PoS blockchains. Our problem is that research rewards are distributed when a user stakes. For this, we propose MRC, BM, and SBD. CBR is not intended to fix the low balance user problem with regards to staking.
I am still uncertain about how you see CBR as a possible disincentive and what the possible adverse economic effects would be under certain scenarios. Could you clarify more please = ). Break down your proposed quadratic curve and give a few scenarios, perhaps.
I will when I get a chance to detail my thoughts! I am at a major fencing tournament today in Kansas City, so it may be later on today/this evening. I am not sure I even agree with my own suggestion. It may be TOO sensitive. Is there a good way to determine the addresses and balances that are NOT staking? What I would like to do is build a detailed model literally address by address from the top down and run some scenarios system wide. This would be far more valuable than just anecdotal examples.
As a related point, the system is currently underpaying interest. (I think the latest figure was something around 0.98% per annum system wide based on the interest paid in GRC divided by the GRC supply), which if I understand things correctly is a symptom of extreme non-participation by some clients (because the interest, while accrued, doesn’t actually get paid until staking occurs). So am I right to say that there are many wallets with accrued interest and high balances that haven’t staked in months/years - that literally amounts to the difference between the actual interest actually paid out and the 1.5% per annum target?
If so, does the network actually owe all this interest to the investors? My guess is that if that is true, some attention has to be paid to the disposition of accrued, but not realized interest during the transition, otherwise people could get upset.
I think the transition period will have to be widely advertised, but in the end if someone is not paying attention of the course of say, 6-12 months, that is their loss. Similar to how the PoW->PoS switch worked, there will be some disappointed parties, but what can you do?
At the same time, perhaps there's a way to snapshot interest owed so when the person does bring coins online they can receive that back interest in their first staked block.
There are potential economic repercussions to implementing MRC/BM/SBD, ie, without the buying pressure from new users purchasing GRC on exchanges for staking purposes would our volume be so low that we would end up getting removed from the exchanges?
This is exactly what is being proposed in CBR. You've independently come up with the exact manner in which CBR is intended to work! More coins = higher probability to stake. Thus whales will stake with higher frequency - still earning de facto interest.
Not quite. My suggestion provides essentially a quadratic curve, because you get the effect of both frequency and balance, whereas CBR, you get only frequency. Maybe quadratic is too much though. The problem with this is to balance all of the incentives/disencentives for the whales, medium, and small balance holders so as not to upend the gridcoin economy in the transition.
Unless I am misunderstanding something, we also still have the problem of long staking periods (low frequency) for really small balance holders, which remains a discouragement for new participants that are not using a significant initial investment.