Bitcoin Classic: a future with multiple "Bitcoins"
Posted: Fri Jan 15, 2016 8:54 am
Recent hype over Bitcoin Classic suggests that a catastrophic decision is about to be made that will have long-term consequences for cryptocurrency. Bitcoin Classic, a temporary solution to a permanent problem, promises to raise the Bitcoin blocksize to 2MB, with no planning or provisions for a future fork. While I've expressed opposition to a one-time fix multiple times before, this article describes why Bitcoin Classic in particular is the worst possible solution to this problem. In fact, it is worse than doing nothing at all.
There are a number of reasons why I think that forks set back a cryptocurrency significantly, even if they do succeed. I talked with Chris and we estimated that only 30% of currencies that have had a contentious fork since we started have resulted in either one of the forks surviving the transition. There is little reason to suspect that bitcoin's odds will be any different in that regard, because the factors that caused those coins to fail are present in bitcoin as well.
Earlier in the year, when I urged Gavin Andresen to select a permanent solution to the blocksize debate, I pointed out that exchanges have the power to commit to a fork, not miners. Because exchanges serve customers, traders are the ultimate arbiters of the decision. If, when a coin announced a fork, every exchange agreed to upgrade to and only exchange the new coin, miners would then have the final say as to whether to go ahead or not. In the real world, however, that doesn't happen. If customers are satisifed with the old fork, then an exchange won't see any reason to cut its revenue by upgrading to something people won't buy. If it were that easy, then miners could again recognize that the fork isn't going to be successful and continue mining the old fork.
But again, that doesn't happen. As has occurred with many altcoins, Cryptsy may trade one fork and Bittrex another. Seeing the competition, the exchanges then add the other market. Once this happens, there are now multiple markets on each exchange running different versions of the same wallet. This situation is currently happening with both contentious forks and also with markets with coins that are being replaced. There is one coin essentially being replaced with a new coin under the same name - where there is supposed to be a fixed exchange rate - but the Bittrex markets diverge from that exchange rate as people arbitrage and speculate on which version will be successful.
The Bitcoin Classic developers are creating a precedent of a temporary hard fork whenever the blocksize reaches a crisis. It is possible, as some believe, that a second hard fork will result in a permanent solution. I see it as far more likely that, should bitcoin survive the first fork, people would assume "that was easy" and the future will be a series of one-time forks every time the crisis recurs. This precedent is disastrous. It would mean that we can expect that, every few months or a year at unexpected intervals, tens of millions of dollars will need to be spent on coding, testing, consumer confusion, lost money due to hacks or not knowing to upgrade, and so on. But while the financial losses in these auxiliary costs would be significant, few are mentioning the market realities of what will happen when these forks occur.
If miners start to mine significant numbers of blocks on Bitcoin Classic, it is likely that many exchanges (especially ones that already list multiple coins like Poloniex and Btc-e) will begin to offer markets in both forks. Each market would be backed by a different daemon, just as altcoins are backed by different daemons. In the beginning, both forks would see the same bitcoins as valid. After the first large block is published, the markets will then be severed. The price of 1MB bitcoins and 2MB bitcoins, summed together, would be some total less than the current price (because fear would cause a market panic to some degree). You'll be able to bet on the future value of these markets even before the seminal block is published, so smart people will sell out the fork they think will go down in value.
People will still want to own stakes in the 1MB netowrk because the value of 1MB bitcoins won't just be theoretical. If theymos and his forums only accept donations in version 1 bitcoins, then people who want to spend money on his services will want to retain 1MB bitcoins and get rid of their 2MB bitcoins. Some percentage of merchants are going to refuse to upgrade - either because they intentionally don't support larger blocksizes, because their customers use 1MB bitcoins, because they didn't know that an upgrade was occurring, or for a number of other reasons. Blockstream may decide to only deploy its solutions on the 1MB network. Core development is likely to continue on the 1MB fork and given how many people are involved in bitcoin, at least one person is going to port changes from the 2MB version to the 1MB version.
The lesson to be learned is this: the 1MB bitcoin network won't just go away. Instead, there will be two markets indefinitely. What is the major reason it won't go away? Because if the 2MB fork is as successful as people say it's going to be, it will become full again. Transaction fees will rise, and the 1MB fork - a well-tested network that was once supported by everyone around the world - would then be empty! The 1MB fork would immediately be far more valuable than any of the other altcoins, because it takes even less effort to convert fee-sensitive services like gambling to the 1MB fork than it does to use litecoin. Not only that, but there would be legacy services that were too difficult to upgrade and for which 1MB bitcoin is good enough. Or, there may be darknet markets that want to stay under the radar from people who datamine the 2MB blockchain.
One of the reasons that it will be extremely difficult to achieve 75% acceptance of 2MB bitcoin is because there is an economic incentive for a percentage of miners to keep mining the 1MB fork, even if it is clear that the 2MB fork is likely to happen. If people believe that 30% of consumers aren't willing or able to come along, then smart miners would refuse to upgrade, since the lower difficulty on the 1MB fork would earn them more profit.
If Bitcoin Classic is adopted, this sequence is likely to happen multiple times. There will be 1MB, 2MB, 4MB, and however many other coins there will be, either ad-infinitum or until people can agree upon a permanent solution. If we make the non-trivial assumption that every fork succeeds, the price of bitcoin will crash every time as smart consumers sell off their coins from the fork they don't plan to use anymore, or simply panic at the prospect of value leaking onto yet another fork. Since forks take time, this process of extreme volatility will last for months. It's difficult to imagine ordinary people wanting to use a system where they have to constantly hedge their bets to retain their savings - as well as distinguishing between all the coins backed by angry proponents that all claim to be "bitcoin." While some people are undoubtedly panicked due to Hearn's departure, others correctly see that there is no way that either fork will remain valued at its current value during a forking process, so they are selling to buy later.
(As a final comment, the choice of "bitcoin classic" is also the worst possible name that could have been chosen. If you don't know anything about cryptocurrency and need to set up a wallet, which coin would you think is more current - the one theymos says is the "core" version, or the "classic" version?)
While there is much blame to go around for Bitcoin Classic, I place the most fault on Gavin Andresen, who has made a critical mistake in diluting his support. He made the correct decision initially - bitcoin XT results in one fork that, because it grows at the upper limits of technology over time, would definitively overwhelm the value of the 1MB fork. Bitcoin XT needs a leader, and Andresen should step in to coordinate development of XT, and stop promoting other solutions that will have extreme unintended consequences. In providing support to blocksizes that will be just below a crisis level, rather than sizes being limited only by what is technologically possible, he ensures that the old forks are going to retain value. It's deeply disappointing that such a visionary leader has failed to see the consequences of one-time forks on the markets.
There are a number of reasons why I think that forks set back a cryptocurrency significantly, even if they do succeed. I talked with Chris and we estimated that only 30% of currencies that have had a contentious fork since we started have resulted in either one of the forks surviving the transition. There is little reason to suspect that bitcoin's odds will be any different in that regard, because the factors that caused those coins to fail are present in bitcoin as well.
Earlier in the year, when I urged Gavin Andresen to select a permanent solution to the blocksize debate, I pointed out that exchanges have the power to commit to a fork, not miners. Because exchanges serve customers, traders are the ultimate arbiters of the decision. If, when a coin announced a fork, every exchange agreed to upgrade to and only exchange the new coin, miners would then have the final say as to whether to go ahead or not. In the real world, however, that doesn't happen. If customers are satisifed with the old fork, then an exchange won't see any reason to cut its revenue by upgrading to something people won't buy. If it were that easy, then miners could again recognize that the fork isn't going to be successful and continue mining the old fork.
But again, that doesn't happen. As has occurred with many altcoins, Cryptsy may trade one fork and Bittrex another. Seeing the competition, the exchanges then add the other market. Once this happens, there are now multiple markets on each exchange running different versions of the same wallet. This situation is currently happening with both contentious forks and also with markets with coins that are being replaced. There is one coin essentially being replaced with a new coin under the same name - where there is supposed to be a fixed exchange rate - but the Bittrex markets diverge from that exchange rate as people arbitrage and speculate on which version will be successful.
The Bitcoin Classic developers are creating a precedent of a temporary hard fork whenever the blocksize reaches a crisis. It is possible, as some believe, that a second hard fork will result in a permanent solution. I see it as far more likely that, should bitcoin survive the first fork, people would assume "that was easy" and the future will be a series of one-time forks every time the crisis recurs. This precedent is disastrous. It would mean that we can expect that, every few months or a year at unexpected intervals, tens of millions of dollars will need to be spent on coding, testing, consumer confusion, lost money due to hacks or not knowing to upgrade, and so on. But while the financial losses in these auxiliary costs would be significant, few are mentioning the market realities of what will happen when these forks occur.
If miners start to mine significant numbers of blocks on Bitcoin Classic, it is likely that many exchanges (especially ones that already list multiple coins like Poloniex and Btc-e) will begin to offer markets in both forks. Each market would be backed by a different daemon, just as altcoins are backed by different daemons. In the beginning, both forks would see the same bitcoins as valid. After the first large block is published, the markets will then be severed. The price of 1MB bitcoins and 2MB bitcoins, summed together, would be some total less than the current price (because fear would cause a market panic to some degree). You'll be able to bet on the future value of these markets even before the seminal block is published, so smart people will sell out the fork they think will go down in value.
People will still want to own stakes in the 1MB netowrk because the value of 1MB bitcoins won't just be theoretical. If theymos and his forums only accept donations in version 1 bitcoins, then people who want to spend money on his services will want to retain 1MB bitcoins and get rid of their 2MB bitcoins. Some percentage of merchants are going to refuse to upgrade - either because they intentionally don't support larger blocksizes, because their customers use 1MB bitcoins, because they didn't know that an upgrade was occurring, or for a number of other reasons. Blockstream may decide to only deploy its solutions on the 1MB network. Core development is likely to continue on the 1MB fork and given how many people are involved in bitcoin, at least one person is going to port changes from the 2MB version to the 1MB version.
The lesson to be learned is this: the 1MB bitcoin network won't just go away. Instead, there will be two markets indefinitely. What is the major reason it won't go away? Because if the 2MB fork is as successful as people say it's going to be, it will become full again. Transaction fees will rise, and the 1MB fork - a well-tested network that was once supported by everyone around the world - would then be empty! The 1MB fork would immediately be far more valuable than any of the other altcoins, because it takes even less effort to convert fee-sensitive services like gambling to the 1MB fork than it does to use litecoin. Not only that, but there would be legacy services that were too difficult to upgrade and for which 1MB bitcoin is good enough. Or, there may be darknet markets that want to stay under the radar from people who datamine the 2MB blockchain.
One of the reasons that it will be extremely difficult to achieve 75% acceptance of 2MB bitcoin is because there is an economic incentive for a percentage of miners to keep mining the 1MB fork, even if it is clear that the 2MB fork is likely to happen. If people believe that 30% of consumers aren't willing or able to come along, then smart miners would refuse to upgrade, since the lower difficulty on the 1MB fork would earn them more profit.
If Bitcoin Classic is adopted, this sequence is likely to happen multiple times. There will be 1MB, 2MB, 4MB, and however many other coins there will be, either ad-infinitum or until people can agree upon a permanent solution. If we make the non-trivial assumption that every fork succeeds, the price of bitcoin will crash every time as smart consumers sell off their coins from the fork they don't plan to use anymore, or simply panic at the prospect of value leaking onto yet another fork. Since forks take time, this process of extreme volatility will last for months. It's difficult to imagine ordinary people wanting to use a system where they have to constantly hedge their bets to retain their savings - as well as distinguishing between all the coins backed by angry proponents that all claim to be "bitcoin." While some people are undoubtedly panicked due to Hearn's departure, others correctly see that there is no way that either fork will remain valued at its current value during a forking process, so they are selling to buy later.
(As a final comment, the choice of "bitcoin classic" is also the worst possible name that could have been chosen. If you don't know anything about cryptocurrency and need to set up a wallet, which coin would you think is more current - the one theymos says is the "core" version, or the "classic" version?)
While there is much blame to go around for Bitcoin Classic, I place the most fault on Gavin Andresen, who has made a critical mistake in diluting his support. He made the correct decision initially - bitcoin XT results in one fork that, because it grows at the upper limits of technology over time, would definitively overwhelm the value of the 1MB fork. Bitcoin XT needs a leader, and Andresen should step in to coordinate development of XT, and stop promoting other solutions that will have extreme unintended consequences. In providing support to blocksizes that will be just below a crisis level, rather than sizes being limited only by what is technologically possible, he ensures that the old forks are going to retain value. It's deeply disappointing that such a visionary leader has failed to see the consequences of one-time forks on the markets.