October 26th 2014 Weekly Hashrate Contributor and Network StatisticsOther Weekly Hashrate Contributor and Network Statistics postsWelcome, miners.
Changelog:
Nil.
Errors:
Discus Fish still seem to update their data irregularly. Until they update more regularly, their data in the second table , and for figure 4, 5 and 7 will only be based on that subset.
Notifications:
Nil.
0. Block propagation latency and sand size.
I don't have anything interesting to write about pools or the network this week. For the last couple of weeks I've been spending all my free time investigating block propagation times provided by coinometrics.com using their network monitoring tools.
I've found that block propagation latency (time since a block was first reported by a peer to a monitor until reported by some other arbitrary peer) is distributed in much the same way as the size of grains of sand, which is fascinating. It implies that block propagation latency distribution is a mixture of many differently distributed (heterogeneous) normal random variables.
The tricky part has been describing the relationship between block propagation latency and block size. There is a clear relationship between the block latency distribution and block size, but there are four parameters that describe the block latency probability distribution, and three of them change as block size changes - attempted to describe the relationship is what has taken so much of my time for the last two weeks.
I think I'm almost done now and I hope to post something soon so I can get out of this rabbit hole and start answering all your emails.
Explanation of the tables and charts.
Table 1: Solved block statistics. This table lists all statistics that can be derived from the number of blocks a hashrate contributor has solved for the past week. Block attributions are from either coinbase signatures, known generation addresses or claimed by a particular pool block history. Includes non-Pool hashrate contributors. Note that actual pool hashrates when derived from shares submitted per unit time will be more accurate than the hashrate estimates given in this table.
"Unknown" is not an entity, but simply the group of blocks to which I cannot give attribution using the methods given above.
Table 2: Pool reported block history statistics. This table lists all statistics that can be derived from the number of blocks a hashrate contributor has solved for the past week using all solved blocks - both valid and orphaned - and difficulty 1 shares per round.
A much more accurate estimate of the hashrate, confidence intervals are unnecessary.
Orphan races lost, and percentage of solved blocks that were not added to the blockchain.
"Luck" is the usual difficulty 1 equivalent shares per round / mining difficulty, or (equivalently) accepted shares / expected shares.
CDF: The cumulative density function (CDF) measures the percentage of the time this number accepted shares / expected shares would be less than the calculated value, given the number of valid + invalid blocks.
Bitcoin per Gigashare. This figure is not an indicator of how much a miner should have expected per one billion Difficulty 1 shares (or one thousand difficulty megashares, etc), since it doesn't take into account the reward method or fees charged. Rather, it should be considered as a "luck" index that also incorporates the number of orphaned blocks and the current reward per block.
Since BTC Guild doesn't report shares per block but does report transaction hashes for all blocks, luck calculations cannot be calculated but orphaned blocks can. Pools such as "Discus Fish" that don't have a public pool interface cannot be included.
Figure 3: Percentage of blocks solved each week for the current top ten contributors.
Data is calculated from the number of blocks each contributor added to the blockchain during the week. The points are the actual data; the lines are exponentiated smoothing splines of the log of the data.
You can view all previous charts at
http://organofcorti.blogspot.com.au/search/label/weeklypoolstatistics and other posts and fun things at
http://organofcorti.blogspot.com. Follow me on Twitter @oocBlog for notification of new posts as soon as I publish.