In my last article I wrote about how BIP-148 adds nothing but noise to forking decisions, and how ‘bit 4’ signaling fails to meaningfully telegraph miners intentions to commit hash power to a fork. But criticisms should never come without offering a solution when you care about the outcome.
One of the great failures of SegWit2X is that they gave away SegWit without any assurance of the 2X. What if, instead of merely putting a single bit in a mined block to signal other miners of their desire to fork, miners had to lock up some bitcoins as a part of their signaling? Would it even be possible for the fork to be canceled?
A staking mechanism has often been suggested as a replacement for mining when it comes to putting together blocks, but these stakes are based on the amount of a base currency held by the miner. As I explained in my previous post, liquid assets invite a certain class of flighty investors. This ability to exit quickly makes their stake less reliable than the investment miners have made into mining, who’s capital is tied up in ASICs dedicated to that coin. However, while staking mechanisms might not be workable when mining regular blocks, they could properly be applied to forking governance. So how would a miner tie up coins that could only be recovered after a hard fork?
Similar to ‘bit 4’ signaling, but instead of setting a bit, you send part of the block reward / fees to a public address. To ensure that the address cannot be spent, we base the address off the hash of the last block mined. The first miner to send a transaction to an address equal to the Merkle root of the last block starts the signaling process. From that point on, each miner will equal or increase the fraction of the block reward they send to that address, and we’ll split the epochs in to two-week intervals. Once the amount of the block reward sent to the Merkle address reaches, say, 80% of the sum of an epoch’s block rewards, then the hard fork is triggered. Now over time, the block rewards sent to the Merkle address will be returned to the miners who sent them.
Now, there is one problem with this scheme. Any miner that locks up funds early on is taking a risk that their funds will be locked forever if the hard fork doesn’t occur. Conversely, any miner not signaling is able to sell their funds immediately. To balance out these asymmetric incentives, any coins mined after the initial signal, and sent to an address other than the initial mining wallet will be distributed the signaling miners. Additionally, the earlier and greater amount a miner sends to the Merkle address, the larger share of those distributed coins they will receive.
Some may view the reversal of miner sold coins during this time period unfair to any potential buyers, but due to the ability to trace any coin, all of this risk can be priced into the coins sold by miners at the time. Not only that, but this signaling can even come after an initial ‘bit 4’ signaling period to ensure all interested parties are informed of the upcoming fork.
These sorts of mechanisms are what I would call “Illiqiud Governance”. Owning an ASIC miner is an illiquid method of creating blocks. Tying up coins behind a future Hard Fork is an illiquid fork governance method. This kind of proof ensures that no hash power is wasted on doomed forks, and that all parties signals are as honest as we can make them. In my next article I will discuss how “Illiquid Governance” can even bring this benefit to non-mining Blockchains, and how other forms of illiquid “assets” can be created around Proof-of-Stake mechanisms.