After previously discussing the biggest parasites in Dogecoin 1.5 in part one, this part will focus on version 1.6 and the changes herein compared to version 1.5.
Similar to before, a sample containing the first 7,500 blocks on Dogecoin 1.6, starting at block 145,001 to 152,500, is used for the following analysis. The results are again separated per wallet address that were rewarded for completing a block. Only addresses that received rewards for more than a 100 blocks are included.
Block rewards
It is no longer possible to observe any differences in block rewards. Even though block rewards were set to random amounts in version 1.5, some pools found out and took advantage of the fact that in reality it was possible to predict block rewards. By targeting Dogecoin at the right time, they succeeded in getting unnatural large rewards on average. Dogecoin 1.6 introduced fixed block rewards to counter this, which would amount to 250,000 DOGE per block at the start.
For the biggest dedicated Dogecoin miner in the network, DogeHouse, this means it will mine about 30% more coins on average, given a roughly equal amount of blocks mined as shown below. This is also the case for other dedicated miners, which were all receiving less than 200,000 DOGE per block on average before the update to Dogecoin 1.6.
Time to solve
The time required to solve a block was another item examined for Dogecoin 1.5 in part one. The block target time remained the same at one minute in both versions. It could be expected that the new difficulty algorithm would have an effect on this. The DigiShield difficulty algorithm, which was introduced along with fixed block rewards, was meant to allow for very fast difficulty adjustments (every 60 seconds). In the previous situation, if a multipool would leave the network the total hashrate would drop immediately, but the difficulty would take some time to adjust. Dedicated miners would be left relatively high difficulties and longer times required to mine a block as a result.
Rapid difficulty adjustments should result in about equal difficulty on average for any time required to solve a block. For better understanding, consider a block with a difficulty of 1000 that would take about a minute to solve with a total hashrate of 100 GH/s. In Dogecoin 1.5, calculation power would drop significantly if a multipool contributing 25 GH/s would suddenly leave the network. The total hashrate would drop to 75 GH/s, yet the difficulty would remain equal. It would then take the network longer to solve the same block. In Dogecoin 1.6, the difficulty will adjust quickly to a lower level, so that any excess times to solve a block are avoided.
It was observed that the algorithm was seemingly working as expected, as the relationship between difficulty and time required to solve a block was strongly reduced. It can, however, be seen that some pools still succeed in solving blocks at an unusually fast rate on average.
The biggest winner
The plot reveals that there are still several addresses that are able to mine many blocks significantly below the target time of one minute. There is even one address in the bottom right corner that is in the second position when it comes to the number of blocks mined, yet it takes the least amount of time to complete block on average (just 32 seconds). The address belongs to Wafflepool, a multi-coin (scrypt) mining pool. This pool is now by far the biggest parasite of Dogecoin 1.6 taking a total of 1,142 out of 7,500 blocks (worth 285 million coins). It is also the main reason for the observed increase in negative times to solve a block. A full table with addressess, (average) coins mined and average times to mine a block can be found here. This includes all data for addresses that mined more than a 100 blocks in either Dogecoin 1.5 or Dogecoin 1.6 (from the sample of 15,000 blocks around block 145,000).
Verdict
In short, Dogecoin 1.6 has benefited dedicated miners a lot, increasing their reward by roughly 30%. The increase is, however, mostly due to the fact that block rewards became fixed. The DigiShield difficulty algorithm succeeds at reducing the relationship between difficulty and times required to solve a block. But this doesn’t automatically mean it also achieves what it was hoped to achieve. It is, unfortunately, not capable of fully restoring balance between dedicated pools and multipools. It did drive away a few pools that were exploiting Dogecoin, but in return a single bigger threat (given the numbers of blocks mined and its involvement in the increase of negative times to solve a block) has taken centre stage.
