Following the report on how the update to Dogecoin 1.6 affected Dogecoin mining in general, Dogeconomist will zoom in on the actual winners and losers of the update. This part will first cover who was taking advantage of the situation in Dogecoin 1.5, and exactly how much this affected Dogecoin.
Multipools and random block rewards
Dogecoin 1.5 had two problems which were both addressed in version 1.6. The first one being the block rewards which miners would receive upon solving a block. Before the update to 1.6 the rewards were set to a random amount between zero and 500,000 DOGE. This set up had been chosen to add an additional fun factor to mining. Unfortunately, it turned out to be possible to predict block rewards. This opened the possibility for so-called multipools to mine Dogecoin if they knew a big reward was coming.
Multipools are “multi-coin profit-switching mining pools” that “use their hashing power to mine whichever coin is most profitable at a given moment.” Coin profitability is determined as “a function of both current price and current difficulty. The idea is to mine coins while their difficulty is low, so that you get more coins for your hashpower.” This is a problem for altcoins in general, and the predictable block rewards made the problem even bigger for Dogecoin.
As of Dogecoin 1.6, block rewards are no longer random. Additionally, the DigiShield difficulty algorithm was implemented. Difficulty adjustment are now “nearly instantaneous, eliminating the problem of large pools surging in when the difficulty is low and abandoning once it adjusts.”
The impact of multipools and predictable block rewards can be easily observed on the blockchain. A sample of the last 7,500 blocks on Dogecoin 1.5, starting at block 137,501 to 145,000, is used for the following analysis (similar to previously used data). The results are separated per wallet address that were rewarded for completing a block. Only addresses that received rewards for more than a 100 blocks are included (leaving a sample of 6,223 blocks).
As block rewards were a random amount between zero and 500,000 DOGE, the expected average on many blocks would equal 250,000 DOGE. It can, however, be shown that several addresses received an unusual amount of DOGE on average.
It can also be observed what the impact was for dedicated miners, as their reward would fall to about 200,000 DOGE on average; 50,000 DOGE less than expected for truly random block rewards. Likewise, some pools were unusually fast at completing blocks as shown below. The target time for one block is one minute.
The multipools were able to get both the best rewards and the best times, which basically created two groups in the network. Dedicated miners were left with not only less rewards, but also longer times before receiving a reward. The split can clearly be seen if the average amount of time to solve a block is plotted against the average reward per address.
A full list of addresses with their average reward received per average amount of time can be found here.