800 billion RMB evaporated in less than a year. Are Ethereum miners still mining?

Introduction: In the past year, the market value of Ethereum has dropped from 125.9 billion US dollars to less than 10 billion US dollars, and nearly 800 billion yuan has evaporated. Under such circumstances, are the miners who maintain the operation of the blockchain still there?

The good news is that Ethereum miners are still there, but they are making much less money; the bad news is that some miners are using certain strategies to gain more profits. In the face of interests, miners have begun to change their strategies to gain more profits. The number of Ethereum empty blocks has skyrocketed 5-7 times since September. Some mining pools use empty blocks as their main profit method, and 86% of the blocks mined in a year are empty blocks.

Original author: Bogdan Gheorghe

Published on: Medium - A Minor Winter, the following is the original translation provided by BlockBeats:

"This is the best of times, it is the worst of times.

This is an age of wisdom, this is an age of foolishness,

This is a time of belief, this is a time of doubt,

This is a season of light, this is a season of darkness,

This is the spring of hope, this is the winter of despair,

People have everything before them, and people have nothing before them."

Although the bear market in the digital cryptocurrency trading market has eased in the past few months, people still believe that there will be a big drop in the next wave, investors and mining farms will leave the market one after another, rumors are spreading, and people are panicking.

However, we (the original authors) noticed from the data that this exit trend is not that severe, and the mining farms have found a way to survive without changing their profits, that is, mining empty blocks. We analyzed statistical data to understand why this behavior is profitable and identified the main players who follow this practice.

Winter is coming? Where will the mines go?

Everyone is aware of the price drop in Ethereum that has been a concern over the past six months. We decided to take a look behind the scenes to better understand the actual impact of what has happened in recent months on mining farms and pits. Let’s go deep into the mines to see what has really been going on in the underground world over the past few months.

The number of blocks remains stable

A high-level view of the network over the past six months shows that the number of blocks has remained at the same level, and the number of transactions processed per day has remained at around 600,000.

But this does not mean that the amount of mining activity remains constant. Ethereum's difficulty adjustment algorithm is constantly adjusting the mining difficulty to keep the block generation time within a stable range of 10-19 seconds. Therefore, when the number of blocks is calculated over time, the block mining rate always remains stable.

Mining activity reduced

At the same time, a drop in difficulty and hashrate also reflects a drop in mining activity—it doesn’t necessarily mean a drop in the number of miners, but it certainly reflects a drop in the number of mining rigs. “Miners” here refers to the addresses that mine blocks on the blockchain, not the actual individuals behind them. In most cases, these addresses are mining pools, where hundreds of individual registered users provide their decentralized hashrate.

Difficulty measures the complexity of the task that a miner needs to solve when mining a block. Hashrate refers to the sum of the computing power of miners across the network. When mining activity decreases, the hashrate decreases, so it takes more time to solve a problem of the same difficulty as before. The unit used to measure difficulty is TH (terahashes, or trillions of solutions). Therefore, the total network hashrate is actually the sum of the hashrates of all miners, which is measured in trillions of hashes per second.

From the above figure we can see that as the computing power fluctuates on the network, the difficulty of task processing will also change over time, with two large drops in September and November. The scatter plot below also shows a similar positive correlation.

The scatter plot of the processing difficulty and computing power of each block shows a clear positive correlation between the two, just as we explained in the previous theory. An interesting finding is that these scatter plots are divided into two clusters, with the divergence of the average difficulty value being around 3200TH. We believe that this result reflects the lag in miners' response to large adjustments in network difficulty, especially during its rising period. From Figure 2, we can see that there are two steep slopes in the processing difficulty (September and June), both of which occurred around 3200TH, while the computing power shown by the orange line did not follow up the adjustment in time and reacted late due to the time lag.

Whenever the average difficulty drops significantly, miners can respond in time and stop contributing computing power when the computing power is sufficient. If the average difficulty increases, the situation will be different. Miners will find it difficult to estimate when this wave of difficulty increase will end, so they may overestimate (appeared in September) or underestimate (appeared in June). There may be more explanations around this pattern, and more hypotheses need further research and verification.

The number of miners (mining pools) remains stable

The chart below shows that the average number of different miner addresses has not changed significantly recently, which may indicate that miners are not responding to the sluggish cryptocurrency market conditions directly, but rather adopting a reduction strategy. This is a good sign, which may indicate that their confidence in the blockchain network has not been shaken.

This is to be expected - after all, they are still in the game, still chasing profits, so by design, mining activity should fall when profit margins fall. Mining equipment requires a lot of electricity supply, which increases the cost of miners.

In addition, when the input-output profit ratio of Ethereum is lower than a certain threshold, the consideration behind it should be this: If I am a miner, paying X amount of US dollars in electricity bills every month to mine 1 Ethereum, if the market price of 1 Ethereum is lower than X, then why should I bother to mine it myself, I might as well buy it directly.

The number of mining individuals (hidden behind mining pools) has decreased

Another good way to quantify the reduction in miners' workload is to calculate the amount of money paid by individual miners who contribute computing power to the mining pool, that is, the amount of money paid by the mining pool to the miners. In the previous chart, we can see that this number has a clear and steady downward trend.

Below are some metrics for the top five mining pools by hashrate, comparing their metrics in the first and second half of the year.

(The original image is too long and has been cut into two parts)

One thing they have in common is that the number of accounts they paid has decreased significantly, with F2Pool seeing the largest decrease.

Let's take Ethermine, the largest mining pool today, as an example. According to the above figure, the number of daily payments is decreasing. The reason why we can see the peak is that Ethermine settles the addresses participating in mining once a week, which increases the number of additional transactions on the payment day by 20%-25%. However, even including those days with higher values, the average payment remains at 0.15-0.16 Ethereum.

Are mining pools mining empty blocks?

Another indicator that can highlight changes in the mining ecosystem is the increase in the number of empty blocks mined recently. Here are a few possible explanations for why miners would intentionally adopt this strategy and the impact it could have.

Is mining empty blocks more profitable? Statistically the answer is yes.

From a miner's perspective, the average total reward a miner receives over a period of time is made up of three different components: block reward (about 89%), fee reward (about 2%), and occasional uncle reward (about 9%). To calculate the average reward a miner receives from each block, the best metric is the block reward and uncle reward (including their respective average fees) - weighted by each miner's respective uncle reward rate (since each miner has a different uncle reward rate, this article uses the average for all examples).

We calculated 3.03 ETH, which is worth $275 at the current ETH/USD exchange rate as of December 11, 2018. This is roughly what a miner gets for each block. We also compared the average block production time and found that for blocks with fewer transactions, the average production time is shorter, with an average block production time of only 9.8 seconds for blocks with less than 10 transactions . Since June of this year, the average production time for all blocks has been 14.5 seconds.

Meanwhile, the 11,741 empty blocks mined during this period took an average of 13.2 seconds.

In addition, due to the shorter time consumption, the security provided by empty blocks to miners is that they have a lower probability of being certified as uncle blocks , but are more likely to become part of the main chain due to the time advantage.

Therefore, considering the profit return of miners, mining empty blocks is more efficient than mining ordinary blocks.

Who is mining empty blocks?

If we look more closely, we can notice that 103 miners have mined these 11,741 empty blocks in the past 6 months.

The top five addresses are:

F2pool_2 (1st — 0x829bd824b016326a401d083b33d092293333a830),

Etherdig (2nd — 0x8d35067233605bef6069191ae0922d134ff80d48),

Ethermine (3rd — 0xea674fdde714fd979de3edf0f56aa9716b898ec8),

Dwarfpool (4th — 0x2a65aca4d5fc5b5c859090a6c34d164135398226)

Nanopool (5th — 0x52bc44d5378309ee2abf1539bf71de1b7d7be3b5)

It’s no surprise that 3 of them (F2Pool_2, Ethermine and Nanopool) are currently in the top 5 mining pools by hashrate. From the table below we can see that F2Pool_2 has shown a significant increase in empty block production. As we have concluded from the top 5 mining pools above, they are also the worst performing miners in the second half of the year.

Digging deeper, the fact that more empty blocks are utilized directly reflects the reduction in average Gas consumption (Note: Gas can be understood as transaction fees in Ethereum). We can plot the gas usage of F2Pool_2 blocks based on the ETH-USD price chart.

The above chart shows a clear change in F2Pool_2's trading strategy, and the accompanying sharp declines in late March, late June, July and September. Now let's compare these prices with the ETH-USD price:

By tracking the biggest downtrends, we noticed that March, July, and September showed the same pattern of declines. So it is clear that F2Pool_2 uses a very price-sensitive trade inclusion strategy.

However, the second-largest empty block mining farm, Etherdig, did not use this strategy to adapt to price fluctuations, and simply mined empty blocks as its main mode. They mined an average of 86% of empty blocks throughout 2018 (compared to 5.5% for F2Pool_2).

However, SparkPool, a role model among the top five mining pools in terms of computing power, has not become an important player in mining blocks.

- END -