Forks will not split Bitcoin

“We can’t hard fork,” the argument goes. “Doing so would result in a situation where there are two ‘Bitcoins.’” Since consumers would not know whether they were transacting with valuable ‘Bitcoin A’ or worthless ‘Bitcoin B’, a dual-coin fork would cause confusion and financial loss to all Bitcoin users. Even worse, it would likely result in a market with two Bitcoins in the future (what a horror!). And it would devalue Bitcoin, causing losses to users. One only needs to look at the Ethereum fork in July 2016 to see that these consequences of a hard fork are obvious and inevitable.
Ethereum’s hard fork did cause devastating losses to some of its community members, with thousands of fans gathering in Shanghai in September to mourn the death of Ethereum.

But Bitcoin is not Ethereum after all, and citing the consequences of the Ethereum fork as proof that the Bitcoin fork will inevitably produce such a result is one-sided and impurely motivated. A very simple example is that since the two Ethereum coins produced after the fork can coexist, why can't Bitcoin fork?

The Ethereum difficulty adjustment period is 1 block (14 seconds)

The Bitcoin difficulty adjustment cycle is 2016 blocks (about 2 weeks)

Here is some necessary explanation, which may be a bit too much for the more sophisticated readers, so I apologize for that: Cryptocurrency mining is like solving a math problem of arbitrary difficulty. In Bitcoin, the difficulty is automatically adjusted every two weeks on average (it may not be exactly two weeks, but it should be every 2016 blocks to be precise - this is the key) to limit the time to solve the problem to 10 minutes. If the mining power of the network increases, which has been the general trend for 8 years, the difficulty will increase, making the problem harder to solve. If the hash rate decreases, the difficulty will decrease. It is this mechanism that keeps the production of Bitcoins at a stable and predictable rate.

Ethereum is different, the average block time is 14s, and it is trending towards 12s. As a cryptocurrency that can be controlled by its chief designer (which is also the key), this difficulty update mechanism can be manipulated like the currency itself. For Ethereum, in the leaderless and decentralized world of Bitcoin, people are more willing to accept it as a scripting language rather than a currency.

This centrally planned model is perhaps the second most important factor in Ethereum’s continued minority status. After a theft of over $1 million worth of Ethereum, the central designers decided to roll back the entire Ethereum blockchain history to return the stolen ether to the victims and prevent the hackers from keeping the bad proceeds. For those who firmly believe that cryptocurrency must be immutable, the act of rolling back severely undermines the most basic characteristic of currency that must be holdable. When early Bitcoin exchanges

When Mt. Gox lost more than 800,000 bitcoins belonging to its users (worth almost $500 million at the time), the idea of ​​rolling back the blockchain briefly surfaced, but was quickly rejected as an action that would undermine the integrity of Bitcoin as a currency.

Those who hold a firm stance on immutability decide that they will continue to use the original (unreverted) Ethereum blockchain, where the victim lost his ether and the hacker kept his loot. Thus, there is user demand for the classic version of the ETC blockchain, but demand alone is not enough, miners are needed to provide security.

Soon after the Ethereum fork, the cryptocurrency exchange Poloniex created an ETC token trading market, giving ETC trading value. Because Ethereum maintains a consistent block time regardless of how much hashrate is created in the network, this means that miners can mine and sell ETC as easily as before, and with the establishment of a triangle relationship between miners, users, and market supply and demand trading platforms, ETC can survive as a branch system of ETH.

This situation will not happen when Bitcoin undergoes an intentional hard fork, and the old blockchain may be retained for a short period of time, but the economic equation of Bitcoin and the economic equation of Ethereum are completely different, so a situation like Ethereum will not exist for a long time.

Following the lead of Bitcoin Classic’s 75% hard fork activation threshold, it seems most likely that Bitcoin Unlimited will fork Bitcoin when there is a significant amount of hashrate support, at least according to ViaBTC founder Haibo Yang:

“First, under what circumstances should a fork be initiated? I recommend following the 75% activation threshold it seems most likely proposed, in other words, the hard fork needs more than 75% of the hashrate in the entire network to support it. If the threshold is set too high, consensus will be impossible because a single opinionated mining pool can veto the change to the entire network. The 75% threshold is more than enough to safely execute a hard fork.”

Unlike Bitcoin Classic, Bitcoin Unlimited does not set an activation threshold for an automatic fork. Instead, once a large enough majority of miners signals that they will accept blocks larger than 1MB, any miner can initiate a hard fork by starting to produce larger blocks. This has two concerns:

This created a fork because other miners and nodes whose software did not support blocks > 1MB ignored these blocks and continued to mine 1MB blocks on the old blockchain.”

Yang also proposed a measure to prevent a premature hard fork: continue to only generate and accept blocks under 1MB until the majority of miners are running Bitcoin Unlimited software. Yang's strategy also suggests that even if enough hashrate support has been obtained, miners should not accept blocks larger than 1MB for two difficulty adjustment periods (nearly a month). This will give users, businesses and Bitcoin Core wallet clients time to reorganize and upgrade their software systems to be compatible with the upcoming fork, so as to avoid those who are after the fork from being self-isolated from the previous network.

Now let’s get back to the difficulty adjustment cycle and why when Bitcoin forks, no one will stubbornly stick to the non-forked blockchain.

Assuming that Yang’s 75% threshold is used as the criteria for forking and that reasonable miners do not switch to a more profitable chain:

Both chains will use the most recent difficulty value and both aim to produce a block every 10 minutes.

The BU chain will have 75% of the hashrate, which means that a block will be found every 13.3 minutes (equivalent to 0.75 every 10 minutes), at least until the difficulty adjusts in 2016 blocks.

The old core chain has 25% of the hashrate, which means that only one block is found every 40 minutes (equivalent to 0.25 every 10 minutes), which is also in 2016 blocks.

Based on this data, we can draw some interesting conclusions. Because a miner's operating costs are fixed (electricity costs are not included in the Bitcoin price), miners on the BU chain spend 33.3% more resources to compete for newly produced Bitcoins, while miners on the old chain need to spend 300% more resources to compete for newly produced Bitcoins.

The field of miners competing for new Bitcoin will be smaller in both cases. With a ten-minute block rate, there will be an average of 144 blocks per day, with a reward of 12.5 BTC per block, for a total daily output of 1,800 BTC. In a 13.3-minute period, the BU chain will find an average of 108 blocks per day, or 1,350 BTC. In a 40-minute period, the core chain can find an average of 36 blocks per day, or 450 BTC.

Based on the current price of $700, we calculated the daily economic return:

1,350 BTC = $945,000

450 BTC = $315,000

The current network hashrate is estimated at 1.91 exahashres per second (1,910 ph/s, 1,910,000 Th/s). To use that as a benchmark, I will be using the world's most widely used Bitcoin miner, the Antminer S7. The power consumption of the S7 is 273w/th, but when considering the PSU that powers the machine, one with an efficiency of about 93%, the power consumption is 294/th or 294kw/ph.

Total electrical energy consumption is 1,910 * 294 = 561,540 kW

The estimated average cost per kWh is $0.035. Most miners in China and Washington State have very low electricity costs. Based on this data, the electricity cost of the equipment driving the Bitcoin network is calculated to be $19,645 per hour.

Then we further calculated that the electricity cost of the Bitcoin network per day is $19,654 * 24 = $471,694.

This is just the electricity cost and does not include the capital expenditure of building a Bitcoin mine.

Using the 75/25 numbers above, the BU chain will cost $353,770 per day to maintain, while the core chain will cost $117,924 per day to maintain. Only when the price of each coin is exactly the same will it make economic sense for miners who stick with the old chain to stay on the original chain. However, it is a foolish idea to expect the price of two coins to remain the same.

At the current rate of 3 transactions per second, Bitcoin’s ability to process 259,200 transactions per day on an already congested network is already remarkable. If blocks on the core chain were to go from 10 minutes to 40 minutes, we’d need 64,800 new transactions. Since the Bitcoin network is already at maximum capacity, reducing capacity by 75% would dramatically exacerbate the problem. Not only does a confirmation take 40 minutes instead of 10, but a block now has to process 40 minutes of network activity, or 7,200 transactions. A 1MB block can process 2,000 transactions, so for every 200 transactions that are processed, there is a backlog of 5,200 transactions that need to be moved to the next block, which takes another 40 minutes to confirm. And so on. Of course, all of this assumes the core chain remains the same as it is now.

At the same time, assuming that the BU chain initially sets the maximum block size to 2MB, the BU chain can process 518,400 transactions per day. Block space will not be scarce, and if the emergency consensus mechanism can be adopted by BU, it means that the size of the block capacity can grow with the expansion of user demand, so that pending transactions can be confirmed within 10 minutes, eliminating the fees spent by users to facilitate transactions.

Let’s compare these two blockchains:

CORE chain: 25% of the market value of the entire blockchain, 0.75 transactions per second, 64,800 transactions per day, 40 minutes of confirmation time. 450 BTC are produced every day. 187,200 permanent transaction squeezes are generated every day. You must bid higher than others to complete the transaction.

BU chain, 75% of the market value of the original unified chain. More than 6 transactions per second, 518,400 transactions per day, 13.3 minutes per confirmation. 1,350 BTC produced per day. Generally no transaction backlog, backlogs generated during peak periods are processed quickly, and if backlogs become regular time, the block size can be further increased. Fair fees, no need to squeeze others out of the line to use the network.

It is easy to see which of the two chains will be given a higher value by the market. However, that is not all: from the beginning of the article let us recall “a simple reason” why Bitcoin will not split.

Bitcoin's difficulty adjustment is not 2 weeks, but every 2016 blocks, with a block time of 13.3 minutes, which means that the BU chain will return to a block time of 10 minutes in 18.62 days. At the same time, the CORE chain needs to run in a paralyzed state for 56 days before a difficulty adjustment is made.

Of course, this all assumes that miners are “stupid” and don’t move to the most profitable chain - and of course, we all know what we are assuming.

Businesses that rely on Bitcoin for transactions cannot afford the expensive market transaction fees of 0.75 transactions per second, and they will naturally tend to choose cheaper, faster and more secure blockchains, which is exactly what Bitcoin has been doing.

It is hard to think of any way to make the value of Bitcoin on the Core chain higher than the value of the coin on the BU chain. As the value of the coin on the Core chain declines, this will further incentivize miners to move to more profitable chains, which will further extend the confirmation time of transactions on the Core chain, reduce transaction throughput, increase transaction backlogs, and extend the difficulty adjustment cycle.

As all of these variables continue to tilt in favor of the BU chain, the total cost of maintaining both chains will continue to cost at least $471,694 per day. It will cost the same amount of money to maintain the hashrate of one chain regardless of whether they are mining on it or not, but one chain will recover quickly and become stronger while the other chain's capacity will continue to slowly decline, just like the challenger, only one can exist.