Hi,
Today I learned how the Nu dividend distribution works. So we take some snapshot of the blockchain at a predefined timestamp and distribute dividends according to the balance distribution at this particular block. Likewise, the blockshares will potentially be distributed in a similar way, such that everyone in the second of the launch of the BKS blockchain will get one BKS for each NSR owned.
Now its not hard to imagine how a system like this affects a market with poor liquidity like NSR. It is to be expected (I would say its certain) that each of these snapshots will trigger a pump and dump because there is an opportunity to buy NSR, take the payout and sell all NSR afterwards. In other words the system increases the value of NSR dramatically for one single block and let’s the value drop back again one block later.
This is not good and instead to incentivize owning a particular number of NSR at a special point in time, I think we should incentivize holding NSR for the amount of time that is used to create the revenue by using shareholder funds as investment (like BKS). This means, we do not take a snapshot, but simply calculate the average balance over the time span of the operation.
Example: A sell side custodial grant passes to sell 1000 NBT within 30 days. Person A owns 50% of all NSR, person B owns 50% of all NSR and person C owns 0% of all NSR. After 15 days, C buys all NSR from A and B.
In the old system, C would either get 1000 NBT or 0 NBT, depending on where the snapshot is. In the proposed system, C will get 500 NBT, and A and B will get 250 NBT each.
I think it is obvious that it won’t make any sense to buy NSR shortly before the payout just to dump them. It also encourages holding your NSR although you know that the auction (or another investment decision) will make pressure on the current price.
Implementation: Not even worth making a pull request, its just this method here: http://pastebin.com/kgXigXwi
Analysis: Let’s consider the last dividend payout of 20,000 NBT by @KTm at 00:01 GMT on 20th of January. The following plots show how dividends would have been distributed when considering different snapshots. In all plots, the X axis is just some address and the Y axis corresponds to the percentage of the total dividends.
In the first plot we compare the dividend distribution if they were distributed one week before the announced date (upper plot) compared to the actual date (center plot). The bottom plot corresponds to the difference between the upper and the center plot.
This peak is an exchange address. It means that about 0.8% were in this address one week before the payout, and were not in this address anymore after the payout. At the same time we see several smaller spikes to the bottom which indicate addresses which got filled with NSR shortly before the payout. Now let’s see what happened to those NSR after the dividend payout by making the same comparison, but using the actual payout and a snapshot one week after the actual dividend:
Many coins were moved back to the peak address above, resulting in a large downwards peak here. Aside from that we clearly see a majority of positive deltas, which indicate that people removed NSR from these addresses.
With the averaged balance system the dividend payout is much less affected:
Comparing plot (2) and (3) reveals that the peak address receives only about 60% of the previous payout. The remaining 40% are distributed among the private keys of the previous owners. There is not much to gain from buying NSR one day before the payout, but a large incentive to hold them for the whole duration of the term.
So I am proposing two things:
- Changing the dividend system to the average balance instead of a single snapshot (this will require a motion, right now I am just asking for feedback)
- @JordanLee, please use this system to distribute BKS if your motion should pass and use the begin of the auction as start date and the launch of the BKS blockchain as end date. I am confident that this will also have a significant impact on the auction price and will reduce the sell pressure on NSR following after the auction.