What does the 75% fork threshold mean?

We know that the 75% computing power fork of Classic has caused a lot of controversy in the circle. Some people think it is too low and it may lead to two chains; some people think it is high enough because the party with less computing power will compromise.

In fact, no one knows whether it is a fork or a compromise by the party with less computing power. Before things actually happen, everything is speculation.

The current fork rule is that if 750 of the last 1,000 blocks are marked with Classic, the fork will be executed. My question is, do we need to have 75% of the computing power to fork? Can low computing power rely on luck to reach the fork threshold?

So, I did some calculations. To be honest, I am not very good at statistics. I kept waiting for someone to calculate, but after waiting for a long time, no one calculated it, so I had to do it myself. Please point out any mistakes.

In fact, this is a binomial probability distribution problem. If the proportion of computing power controlled by Classic is p, then the probability that it will produce 750 blocks in the last 1000 blocks is:

X(750) = C(750, 1000) * P ^750 * (1-P) ^ 250

Then, the probability that it will produce >= 750 blocks in the last 1000 blocks is:

X = X(750) + X(751) + X (752) + ……. + X (1000)

So, I calculated the thresholds 950, 900, 750, and 500 respectively, and the results are shown in the table below.

This table means that if the threshold is 950 blocks, then if you have 95% of the computing power, you will have a 53.75% chance of success, which is very high and will basically succeed. If you have 94% of the computing power, you will also have a 10.06% chance of success, which will definitely succeed after a period of time. If we believe that only one in ten thousand means no chance of failure, then if the computing power is less than 91.94%, it is basically impossible to succeed.

Similarly, if the threshold is 900 blocks, it is impossible to succeed if the computing power is less than 86.04%. If the threshold is 750 blocks, it is impossible to succeed if the computing power is less than 69.64%. If the threshold is 500 blocks, it is impossible to succeed if the computing power is less than 44.10%.

Have you discovered anything? Yes, the higher the threshold, the less luck is involved. If the threshold is 950 blocks, you must reach 92% of the computing power to rely on luck. If the threshold is 750 blocks, as long as the computing power is 70%, you can rely on luck. One day, you may succeed in counterattacking.

Don’t think that luck is unimportant. Do you remember the sudden explosion of 48 blocks in 4 hours some time ago? Some people calculated that it only happens once every few decades.

I only listed the results above and don’t want to draw conclusions. Everyone has their own explanations, and some people don’t care about the fork at 70% of the hashrate, so there is no point in saying more.

Finally, I want to talk about why I do this.

There are risks in doing anything, but we can always reduce the risks through some calculations or simulations. If you really want to promote expansion, do what you can to make everyone more and more clear about the various risks.

I have to say that Classic is very irresponsible . Why? When the consequences of something are unclear, if you must push it forward, you have to assess the risks of the matter.

How to evaluate? In fact, there are many engineering methods. For example, establish a similar network (such as a certain altcoin), call on everyone to build nodes together, and do a fork experiment. Randomly select a certain proportion of nodes to refuse to update, and update other nodes to see under what circumstances there will be a fork and under what circumstances there will not be a fork, and through analysis, establish some feasible mechanisms or processes to minimize the risk of forks. If you think that the cost of real experiments is too high or difficult to implement, you can also do simulation experiments on the computer to see what will happen. After the experiment is completed, publish a complete experimental report and an assessment of various risks, including probabilities and consequences. This is a responsible attitude.

It doesn't mean that risky things can't be done, but it means that you should evaluate the risks as much as possible and minimize them. It's not just a simple sentence "the party with less computing power will compromise" and then start an endless quarrel. I want to support you, but unfortunately I can't convince myself.

On the other hand, the SW solution promoted by Core has already started testing on test-net. This is a responsible attitude. No matter how long the test takes, no matter how many bugs are found during the test, it is a responsible performance. Because the longer the test time and the more careful the test, the more confident we are. The so-called large code changes are not a reason to oppose SW at all. Many software have been torn down and rewritten during the development process. As long as the new version has been tested carefully enough, there will be no problem.

Stop saying that capacity expansion is urgent. If you are really in a hurry, shouldn’t you have done simulation or emulation experiments long ago? I really don’t feel anxious. People who are really anxious will not do no testing.

As for political matters, I want to ask you, if you support Classic, will you no longer be a dictator? Judging from Gavin's performance, he will be more dictatorial than Core. When he was a dictator, why didn't he mention the Garden Theory? After he stepped down, the Garden Theory came out, haha!

Developers can't do bad things. The client needs the support of miners and nodes to implement. If miners and nodes refuse to support, developers can't do anything. The design of Bitcoin is perfect. Developers and miners, no one can easily change the network, and they check and balance each other. This inability to be easily changed is the fundamental reason why our users dare to use it to store value.

From: Coin Technology