Modeling (Social) Consensus About (Cryptographic) Consensus Change

OR: Bitcoin Has Ossified, Now Go Use Decker-Wattenhofer

Every 4 years, Bitcoin has an event called the “halvening”. This causes the transient supply to be halved.

Since the price of a good is a function of the transient demand over the transient supply, and there is a continuous transient demand due to HODLers habit of DCAing into Bitcoin, the Bitcoin halvening causes the price to increase dramatically.

The increase in price catches the attention of larger groups of people, who have not been exposed to the ideas of Bitcoin. Some part of these new people then become HODLers as well, and start buying Bitcoin. Some part of these new HODLers continue on to run fullnodes and have great economic power in consensus of the coin — specifically, the ability to sell off coins in an undesired chainsplit whose rules they dislike, crashing the value of the coins in that chainsplit and disincentivizing miners from mining that chainsplit, thus removing security and usability of coins in that chainsplit.

Now, every change in the blockchain consensus rules must, by necessity, require EVERYONE to agree to it, that is what consensus means. A sufficiently large quorum (measured by economic power, i.e. ability to dump coins of undesired rulesets) can forcibly evict a minority that desires a different ruleset, and declare a slightly-smaller consensus, but due to economic disparity, their preferred consensus rules dominate and the evicted minority either becomes even less powerful economically, or capitulates and converts their coin to the greater consensus (or else suffer by having to always exchange for coins of the greater consensus in order to transact with more people, thus always paying for exchange spreads).

The consensus effect, and the halvening period, combine to the following second-order effect:

  • After every halvening, a new group of users joins the consensus.
  • The new group of users must be convinced of the value of future consensus changes, or else they would need to be evicted (right after joining!).
  • We do not want to evict the new users (as that inhibits hyperbitcoinization).

If it took you more than one halvening period to convince existing users, then the new users after a halvening will also need to be convinced, and you took more than a halvening period to convince the previous users, it seems likely you need to convince the new users a similar time period (yes, you might have learned lessons from explaining to the previous existing users — but the new users are new and you may need to learn new lessons on explaining and justifying consensus changes this time, so you should allocate a similar amount of time as convincing the previous users — you are older and wiser now, but you are older and less energetic and more likely to just say “eh, screw it”).

Thus, I postulate ZmnSCPxj’s Law Of Ossification:

If a consensus change takes more than one halvening period to deploy, it will not happen.

1 Like

Let’s recall very quickly Karl Popper theory of falsification. A theory to be considered scientific, it must be able to be logically proven false by an empirical test.

About ZmnSCPxj’s Law of Ossification, let’s try it with an empirical test.

The first time the merkelized abstract syntax tree (MAST) was mentioned was in 2013. The first time this idea when to birth in some consensus rules (BIP341) was in 2021. There has been a span of 8 years.

All things equals, an halvening period should happen every nSubsidyHalvingInterval.

1 Like

I have a different theory for ossification and its not based on halving: https://www.cia.gov/static/5c875f3ec660e092cf893f60b4a288df/SimpleSabotage.pdf

This is not a troll post either. Some really powerful people who control the real world are involved in bitcoin and do influence things. This includes investors trying to scale bitcoin with custodians as well.

There are various points in this that I wouldn’t agree with, but most importantly, if the model was good at describing reality, then what we should be seeing is that long-term users would agree on consensus change, but progress isn’t made because newer participants to the space block it.

In my opinion, this is not at all what’s happening these days - the disagreement, in particular regarding which covenant softfork to activate (if any) doesn’t seem to be related at all to the time the relevant parties entered the space.