On the consensus rules of the Bitcoin system

Conflict of interest: In addition to RMB, I only hold Bitcoin, and I have a large position in it and continue to increase my holdings. I do not have any altcoins, two tokens, or eight tokens, and I will never have any.

(Jiang Zhuoer likes to write about conflicts of interest in his articles, which seems to indicate that he is a die-hard Bitcoin fan. Then I would also like to state my opinion that I do not own Litecoin. Haha!)

1. Introduction

Bitcoin is a decentralized system ^[1] . The decentralized nature of Bitcoin means that there is no absolute authority in the Bitcoin system, making it particularly difficult to upgrade the Bitcoin system.

Of course, easy system upgrades are not necessarily a good thing. For example, many altcoins often upgrade by hard forking, whether it is to tamper with data or modify rules. As long as the founder raises his hand, it is not a big deal to fork twice a day. On the contrary, it is very difficult for Bitcoin to fork softly or hard.

Bitcoin is progressing extremely slowly. Many people mistakenly believe that this is a bad thing, and worry that Bitcoin will be upgraded too slowly and surpassed by other altcoins. However, Bitcoin still firmly maintains its leading position. Some people also believe that the harder it is to upgrade Bitcoin, the more decentralized it is. In the field of virtual currency, nothing is more valuable than decentralization, so Bitcoin naturally dominates the virtual currency list.

But does the Bitcoin system really not need to be upgraded? Of course it does. Bitcoin still has a lot of room for improvement, and everyone in the system still hopes that Bitcoin can continue to improve. Then, the most important question is, what are the upgrade rules of the Bitcoin system? Who decides? How to decide? How to implement? Satoshi Nakamoto's white paper did not discuss this issue.

I personally think that many people (including a few veteran Bitcoin players) still lack understanding of this issue. This article will discuss in detail my personal understanding of the Bitcoin system upgrade rules.

2. How to upgrade the Bitcoin system

The Bitcoin system upgrade rules determine whether soft and hard forks can be implemented, and how to implement them. They are the most important rules that the Bitcoin system should clarify. However, this issue is rarely discussed. Recently, MAbtc explored the Bitcoin consensus system and emphasized the dangers of hard fork upgrades ^[2] . This idea has caused many people to be afraid of hard forks and believe that Bitcoin can never be hard forked. Chang Jian also wrote an article specifically criticizing Ethereum for hard forking to tamper with blockchain data ^[3] .

Some others completely ignore the Bitcoin consensus rules, believing that a hard fork can be performed as long as it can avoid the creation of two chains. For example, Jiang Zhuoer published two articles discussing how to kill another chain ^{[4] [5]} . In layman's terms, Jiang Zhuoer has no intention of reaching a consensus with others, and even has no concept of consensus. He only thinks about how to kill all dissidents.

I personally think that both of these ideas are going to extremes. The upgrade of the Bitcoin system must be completed through consensus, and it can definitely be completed through consensus. Whether it is a soft fork or a hard fork, there is no problem under the premise of complete consensus. I am not against hard forks, but at this stage, I am against hard fork expansion. This is because there is no complete consensus on hard fork expansion. This issue will be discussed later. The core reason why Ethereum failed in the first fork was that it did not reach a complete consensus. But the recent second fork, because it was to fix a bug, even the original chain ETC must be fixed, so this fork is a complete consensus.

3. Why 95% represents complete consensus

What is full consensus? If everyone in the system agrees on something, then we call it full consensus. Full consensus is different from voting, which ignores the opinions of the minority and enforces the majority rule. However, voting does not work at all in the decentralized Bitcoin system. Ignoring the opinions of the minority will inevitably lead to two chains, and Ethereum is a precedent. If Bitcoin adopted the strategy of majority rule, it would not have developed to this day.

So the question is, in a decentralized system, everyone has different values, experiences, and wealth, how can we reach a complete consensus? This is indeed the case. If the population conforms to the normal distribution, then there will be no 100% consensus on any issue. If we define complete consensus as 100% consensus, then in theory the Bitcoin system will never be able to upgrade. Therefore, this definition is not operational, and we can only settle for the next best option and find the closest solution.

In statistics, we call events with a probability of less than 5% as low-probability events. This means that as long as the consensus reaches 95%, we can assume that the probability that the system has not reached consensus is 5%, which is a low-probability event. Therefore, in the Bitcoin system, we approximately believe that 95% represents complete consensus. Moreover, we also usually use 95% confidence to express measurement errors.

This number cannot be changed to 90% of the 1992 Consensus, nor can it be changed to 75% of Classic, nor can it be changed to BU’s no setting. Because this number has nothing to do with whether to fork two chains. Perhaps a 90% fork threshold can avoid two chains, but it ignores the 10% probability of opposition, and 10% is obviously not a small probability event, so it cannot be counted as a complete consensus.

The next question is, how do we know that the system has reached a consensus of more than 95%? This is the origin of the computing power "voting". In fact, this is not a real vote, it is just to measure the consensus of the system. Only when the measured consensus reaches 95%, can the system be considered to have reached a complete consensus, thus activating the upgrade plan. Otherwise, if the consensus still does not reach 95%, then the entire Bitcoin system will choose not to upgrade and continue to execute the original rules.

4. The Credit of Bitcoin System

In other words, the Bitcoin system respects minority opinions. Even if the minority is only 6%, the entire Bitcoin system will follow the minority and not upgrade. This is the fundamental reason why the Bitcoin system upgrades slowly, but this is precisely why Bitcoin is more trustworthy than all other altcoins. We call it Bitcoin's credit.

Bitcoin credit is especially important to depositors. We depositors rarely care about the progress of Bitcoin because we don’t have time to read Bitcoin news and study Bitcoin dynamics every day. Our biggest wish is to hoard Bitcoin and sleep. Therefore, we hope that Bitcoin upgrades will be completed under complete consensus, so that we can solve the problem once and for all without worrying about forks or additional issuance. After 10 or 20 years, the Bitcoin in hand will still be the Bitcoin when we bought it. This kind of stability is a kind of credit, and maintaining this credit is the most important thing for the Bitcoin system. Any other things are not worth mentioning.

We all know a coin that was once very popular, but later became notorious. This coin is said to have many awesome features, but unfortunately, it has experienced countless hard forks and inevitably increased issuance (when someone has the power to hard fork, he will inevitably increase issuance, just like the central bank). But this kind of thing is absolutely impossible to happen in the Bitcoin system. It is basically impossible for anyone to propose a hard fork plan to reach 95% support. At present, there is only one situation that will execute a hard fork - a serious bug in the system, just like the second fork of Ethereum today.

5. Determine the intentions of each party

Now that we understand the full consensus mechanism of the Bitcoin system, we can judge the intention of each proposal.

Anyone who insists on the 95% fork threshold is seeking consensus. For example, Segregated Witness. Although the Via mining pool has clearly stated that it is opposed to Segregated Witness, the Core team still sets the fork threshold at 95%, which shows that Core's intention is to seek consensus. It would rather this plan not be activated in the end than actively destroy the consensus mechanism of the Bitcoin system. At the same time, it also stated to everyone that developers only release code, and whether it can be activated is decided by users. The rumor that Core controls Bitcoin is self-defeating.

On the contrary, those who do not accept the 95% fork threshold have the primary goal of promoting their own solutions. For example, XT-Classic-Unlimited has been lowering its fork threshold. This shows that they are afraid that their solutions will not be activated, so they are willing to destroy the consensus rules of Bitcoin. Obviously, their purpose is not to seek consensus, but to promote their own solutions by any means. In order to implement their own solutions, they are even willing to forcibly kill others.

I think it is clear who is trying to manipulate Bitcoin. In my mind, the XT-Classic-Unlimited supporters are completely untrustworthy. They trample on the rules at will, and if we give them power, they will be courting death.

In this matter, the most disappointing person is undoubtedly Gavin. Although no matter what he does now, it cannot erase his previous contributions. However, he made many stupid moves on the issue of capacity expansion. In order to promote his own plan, he did not hesitate to break the consensus rules first, and even tried to force a hard fork without consensus. However, the Bitcoin system is very powerful, and even if Gavin is as influential as he is, he will be ruthlessly abandoned by the system as long as he tries to break the consensus rules.

I personally fully support the consensus rules of the Bitcoin system. Although, I like Segregated Witness. However, if more than 6% of the computing power is against it, I will respect other people's choices, because the rules of the Bitcoin system are like this. In layman's terms, I would rather give up my favorite solution and maintain the consensus rules. If Segregated Witness is not activated in the end, I will definitely be very unhappy, but I will still choose to accept it.

The only thing I cannot accept is the trampling of the consensus rules of the system, because it will destroy Bitcoin sooner or later. Kill 50% of the dissenters this time, and kill another 50% of the dissenters next time, and the next time... Bitcoin will be yours.

6. Hard fork is completely free

The conditions for full consensus are too harsh. If someone thinks his proposal is good but cannot get 95% support, does it mean that this proposal has no hope? Of course not, because hard forks are completely free.

If XT-Classic-Unlimited supporters firmly believe that large blocks are right and can reach full consensus, they should set the fork threshold at 95%. On the one hand, insisting that their own solution is the best, on the other hand, constantly lowering the fork threshold is a contradiction in itself. Lowering the fork threshold itself is a manifestation of lack of confidence.

If you are worried about not getting full consensus, the right thing to do is to create a fork of your own, but this coin is not Bitcoin, because every upgrade of Bitcoin is the result of consensus, and non-consensus forks are all called altcoins. Altcoins are not without a future, and the forked chain of Ethereum has surpassed the original chain. So don't get hung up on your name, the market can make the choice, please implement the Unlimited fork immediately.

I really can't understand the supporters of XT-Classic-Unlimited. For example, Jiang Zhuoer, who is the loudest voice, never switches his computing power to Classic, which makes people doubt his motives. For another example, he recently proposed the so-called "safe hard fork", but if you ask where is his code, he will tell you that he doesn't have it. He supports Bitcoin on the surface, but holds a lot of Litecoin. He supports hard fork expansion on the surface, but mines Core. He shouts for decentralization on the surface, but tries his best to kill dissent.

7. When will the hard fork be implemented?

It certainly won’t be now, because the expansion problem is not worth using a hard fork. In the absence of complete consensus, it is not worth the loss to insist on doing so and rashly destroy the long-established consensus rules and undermine the credibility of Bitcoin. Compared with Litecoin, our understanding will be clearer. Its block is equivalent to 4M every 10 minutes. If the public likes large blocks and cheap transaction fees, why don’t the public use Litecoin? Do we really need to kill dissent through a hard fork and then turn Bitcoin into Litecoin? Jiang Zhuoer is not worried, because even if Bitcoin dies, he still has Litecoin. I am different. I am really worried!

However, I don’t think that hard forks must not be implemented. If a major bug is found in the Bitcoin network, the Bitcoin network will face the risk of death if a hard fork is not used. At this time, it is completely feasible to increase the capacity limit while fixing the bug with a hard fork. Since the Bitcoin network hard fork will inevitably reach a complete consensus to fix the major bug, it is completely acceptable to take this opportunity to increase the capacity moderately to 2M or 4M.

8. Summary

(1) It is impossible for a decentralized system to achieve 100% complete consensus, but Bitcoin requires complete consensus for upgrades, so it uses 95% support to represent complete consensus, and considers the probability of 5% opposition to be a small probability event that can be ignored. This number cannot be changed to 90% of the 1992 Consensus, nor can it be changed to 75% of Classic, nor can it be changed to no setting of BU.
(2) The Bitcoin system can only be upgraded when full consensus is reached. Otherwise, the Bitcoin system remains the same. This has become the credit of the Bitcoin system and the basis of its stored value.
(3) The Segregated Witness insists on the 95% fork threshold, which shows that the Core team is seeking consensus. They would rather their own plan not be activated than not abide by the consensus rules. The supporters of XT-Classic-Unlimited continue to lower the fork threshold, which shows that they are just promoting their own plan. The essence of Jiang Zhuoer's plan is to eliminate different opinions, not to seek consensus.
(4) Even if full consensus is not reached, a hard fork can be implemented. Hard forks are everyone’s freedom. Creating a copycat coin does not necessarily mean it will lose to Bitcoin. The most important thing is to keep your word and don’t talk nonsense.
(5) Hard fork expansion will definitely not happen in the near future because it is not worth it. If a hard fork is necessary to solve a major bug in the Bitcoin system in the future, the capacity limit of the Bitcoin block can be increased at the same time.

9. Reference

Notes (↵ returns to text)

1. Bitcoin White Paper: A peer-to-peer electronic cash system
   http://www.8btc.com/wiki/bitcoin-a-peer-to-peer-electronic-cash-system↵

2. Hard Forks and Consensus Networks: Meta-Problems and Limitations
   http://www.8btc.com/hard-fork-and-consensus-network↵

3. Ethereum’s “Ship of Theseus”
   http://www.8btc.com/ethereum-the-ship-of-theseus↵

4. Calling on AntPool to promise to cancel forks below 25% and maintain the unity of Bitcoin
   http://www.8btc.com/kill_small_fork↵

5. [Completely end the debate between soft and hard forks] A safe hard fork is a soft fork expansion
   http://www.8btc.com/safe_hardfork_is_softfork↵