Skip to content
Create pie charts to see the most dominant miners over the history of Bitcoin.
Python Shell
Branch: master
Clone or download
Latest commit 10b88ea Oct 28, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
images increased diagram sizes Aug 30, 2019
LICENSE first commit Aug 7, 2019
README.md changed link Oct 28, 2019
bitcoin-iterate.sh fix - forgot to specify full path to data/txs_scriptsigs_addresses.tx… Oct 28, 2019
gifsicle.sh first commit Aug 7, 2019
known_miners.json GHash.IO address Aug 8, 2019
main.py tiny edit Aug 8, 2019

README.md

Bitcoin Mining Distribution

This code allows you to identify who mined the blocks in your blockchain, and plot pie charts so that you can see the most dominant miners over the history of Bitcoin.

EXAMPLE: https://twitter.com/in3rsha/status/1159468260661895169

In other words, this tool shows you how close miners have come to controlling 51% or more of the blocks in any given 2-week period.

TIP: Miners are identified using the coinbase signatures in known_miners.json

Install (Linux)

Just clone the code.

git clone https://github.com/in3rsha/bitcoin-mining-distribution.git

Dependencies

1. bitcoind

This tool uses your local blockchain as its source of data. So if you're not running bitcoin already, install bitcoind and run it for a few days to download your copy of the blockchain:

sudo apt install bitcoind
bitcoind --daemon

2. bitcoin-iterate

Next, we use bitcoin-iterate to extract the neccessary block data from your blockchain files (in ~/.bitcoin/blocks):

git clone https://github.com/rustyrussell/bitcoin-iterate.git
cd bitcoin-iterate/
make
sudo mv bitcoin-iterate /usr/local/bin/
rm -rf bitcoin-iterate/

3. Python + Matplotlib

We then use a Python script to decode this data and create pie charts using Matplotlib. You probably have both of these installed already.

NOTE: The Python script in this repo is not currently compatible with Python 3.

4. Gifsicle

Lastly we create a cool animated gif of these pie charts using Gifsicle:

sudo apt install gifsicle

Usage

Make sure you have bitcoind running (so that bitcoin-cli is available), then run these scripts in the following order:

# 1. Get the blockchain data we need, which will get stored in a data/ directory.
bash bitcoin-iterate.sh

# 2. Run the main script to identify who mined each block, and save the pie charts to the charts/ directory.
python main.py

# 3. Create a gif from the individual pie charts.
bash gifsicle.sh

After running each of these scripts you will have:

  1. data/all.csv - Block and coinbase transaction data for every block.
  2. charts/*.png - Individual pie charts for every difficulty period (roughly 2 week intervals)
  3. mining-distribution.gif - An animated gif of the entire mining distribution history.

NOTE: The individual pie charts of mining distribution are created for the blocks mined during a target adjustment period (every 2016 blocks). When the target adjusts, the counters are reset.

How does this work?

This tool gathers block data from your local blockchain, using bitcoin-iterate to grab the data and bitcoin-cli to decode it.

It then compares the data inside each block against a list of known_miners.json to figure out who mined them.

How can you identify who mined a block?

When a bitcoin miner constructs a block, they place a special transaction at the top of the block called a coinbase transaction. This transaction is special because it allows a miner to send themselves a fixed amount of bitcoins that did not already exist (called the block reward).

However, whereas in a typical transaction you have to select and unlock some existing bitcoins, this does not happen in a coinbase transaction. As a result, every coinbase transaction has some "redundant" space where the "unlocking code" for existing bitcoins would have normally gone.

This free space is called the "coinbase", and miners can put up to 100 bytes of any data they like in to it. For example, Satoshi put the following in to the coinbase of the first block:

The Times 03/Jan/2009 Chancellor on brink of second bailout for banks

However, most major miners place their own unique text in the coinbase as a way of "signing" the block (figuratively, not cryptographically). Thanks to this data you can ascribe most blocks to a specific miner or mining pool.

Here are some more examples:

0      [...]The Times 03/Jan/2009 Chancellor on brink of second bailout for banks
233030 [...]BitMinter[...]
280806 [...]Happy NY! Yours GHash.IO
294664 [...]Mined by BTC Guild[...]
331672 [...]NOTHING BIG CAME OUT OF SOMETHING SMALL
362391 [...]New Horizons is 23267747 km to Pluto.[...]
377133 [...]Mined by AntPool[...]
416236 [...]Sun Chun Yu: Zhuang Yuan, will you marry me?[...]

NOTE: Sometimes you need to look at the destination address in the coinbase transaction to identify a miner. For example, GHash.IO used the address 1CjPR7Z5ZSyWk6WtXvSFgkptmpoi4UM9BC to collect the block reward, but did not always put their name inside the coinbase.

Why are these charts interesting?

Because if someone is mining 51% or more of blocks, they have the ability to "undo" transactions that have been written to the blockchain.

This is known as a 51% attack.

In short, the blockchain is just a series of blocks built on top of each other, and these blocks contain transactions. Everyone agrees that the "correct" version of the blockchain is the one with the longest chain of blocks. Therefore, if you wanted to remove a transaction from the blockchain, you would need to rebuild a longer chain of blocks to replace the current one.

However, because anyone in the world can become a miner, and every miner will naturally want to build upon the longest chain of blocks (to claim valid block rewards), it's difficult for any individual to out-mine the network and build a longer chain of blocks to replace the existing one.

However, if a single miner has the ability to mine more blocks than the rest of the network combined (i.e. over 50%), then they then also have the ability to build longer chains that other nodes will adopt, therefore "undoing" blocks that have already been mined. In other words, they have the ability to create chains that remove blocks and transactions from the blockchain.

Now, as bad as this sounds it isn't necessarily the end of the world, because the incentive to continue to build upon the longest chain and claim block rewards is probably greater than the benefit of "undoing" a transaction (not to mention the effect it would have on the current value of bitcoin). Nonetheless, the risk would still be there, which is why it's preferable that we see that no single miner is mining more than 50% of blocks at any given time.

FAQ

What do the colors mean?

Nothing at all. I just gave each miner a random color (by hashing their name and taking the last six characters as a hex color value).

It's just a coincidence if two miners are a similar color.

How reliably have you identified who mined each block?

Not 100% reliably, because miners can put whatever data they like in the coinbase field.

For example, a miner could remain somewhat anonymous by not putting any text in there at all, or they could put in the name of a completely different miner (if they wanted to).

With this tool I'm just working with whatever data has been put in to the coinbase, so it's not an exact science.

Wasn't slushpool the first mining pool?

Yes, but they didn't mark their blocks at the start.

This tool only looks in the coinbase field to identify miners, which is why Eligius shows up as the first mining pool instead.

Links

I identified each miner's unique coinbase signatures by browsing through the coinbase transactions manually, which was great fun. These resources were also helpful:

If you're interested in checking out coinbase messages, I display them at the top of each block in my blockchain explorer:

Here's some further information on coinbase transactions and mining:

And some other interesting links:

You can’t perform that action at this time.