What’s going to happen when bitcoin forks (again)?

The scaling debate has been raging for years, and now the bitcoin network will fork again on block 494,784 (around November 16th) as a result of the famous “New York Agreement”.

The bitcoin community experienced its first hard fork when Bitcoin Cash (BCH) launched on August 1st. That was a monumental event in bitcoin’s history that may have contributed to the BTC price promptly shooting up from ~$2,800 USD to ~$4,800 (on August 31st):

Hard to complain when the price does that!

This time, the situation is different.

Bitcoin Cash Hardfork

When BCH launched, it took the community somewhat by surprise. Due to the last minute nature of the launch there was some uncertainty about what the future held. Bitcoin Cash uses the same address format (1aBcD… for single key addresses), shares the same transaction history (and thus balances) at the time of the fork, and has very similar software, so it looked and felt a lot like regular bitcoin.

The big difference between then and now is that:

1. Only a tiny minority of the community supported Bitcoin Cash. Regardless of what you think about the block size, segregated witness has a ton of benefits, and Bitcoin cash went out of their way to remove that code (as well as RBF). Bitcoin Cash was designed as an alternative to those who don’t like the direction regular Bitcoin has gone.
2. No major players were arguing at the time that bitcoin cash was the real bitcoin, though Roger Ver’s bitcoin.com did recently make that claim. It was listed on exchanges as BCH. It’s worth noting that there have been some disagreements about the name “Bitcoin Cash” and “bcash”.
3. Bitcoin Cash implemented something called strong 2-way replay protection, which means that transactions that occurred on the Bitcoin network could not be replayed on the Bitcoin cash network, and vice-versa.

Bitcoin cash wasn’t directly trying to replace Bitcoin, but instead they were offering a different path. The outcome of this is that both chains could coexist in peace, which is roughly what has happened. Bitcoin Cash can now implement all kinds of changes that it thinks best, without having to get approval from the greater bitcoin community, and this is exactly what they’re doing.

I would argue both sides seem happier after the split vs before.

Replay Protection

The Segwit2x hard fork (or what I’ll call B2X moving forward) does not have strong replay protection. Originally, it had no replay protection, then Jeff Garzik added opt-in replay protection (good explanation here), which he later removed after Bitcoin Core developers found a security vulnerability. Garzik currently has another proposal to add only weak replay protection to B2X. This is by design:

Who is doing the“breaking” depends on which you think is the “real” bitcoin.

This means that when transacting on one blockchain, end-users will have to take extreme caution to be sure they don’t accidentally lose their funds on the other. For example, Bittrex’s current cross-chain policy to only refund transactions that exceed $5,000 USD (while charging a 0.1 BTC fee), and that’s actually quite a good outcome since you may eventually get most of your money back. Every bitcoin service will have to scramble to write custom accounting and transaction management software to handle this, though at least they’re well positioned to do so. Tensions have run so high that one software engineer doing that recently got SWATted and posted this:

Standard bitcoin software engineer desk configuration. Using a mousepad with an optical mouse is a bizarre setup, but who am I to judge?

Even without strong replay protection, it will be possible for savvy wallets users to construct transactions that can’t be replayed on the other chain. This is done by ensuring that your transaction contains an input that traces a part of its lineage to a block reward transaction after the hard fork. That way, the transaction would be invalid on the other chain. Managing this process is very difficult, especially for light clients who have limited access to blockchain data.

Investors

Investors are what give the bitcoin network value. Like gold, bitcoin has almost no intrinsic value and is only worth what someone else is willing to pay. Investors currently value Bitcoin Core at a whopping ~6x more than B2X:

You’re going to need a big tinfoil hat to explain this away

While prediction markets are imperfect (due to counterparty risk, liquidity, biased contracts, etc), they are the best forecasters we have. Investors in multiple prediction markets have overwhelming consensus that Bitcoin Core will be worth much more than B2X.

Miners

Currently, ~85% of bitcoin blocks are signaling that they like to see the launch of Segwit2x. Keep in mind that signaling is non-binding, and miners are known to signal for software they don’t actually run. Historically, miners across all blockchains have always followed profitability:

The peaks in BCH mining are miners hopping over from BTC to BCH and back after each BCH Emergency Difficulty Adjustment (EDA) kicks in (source).

Except for cases of extreme ideology, miners point their hashpower wherever it is most profitable to do so. Running their equipment incurs variable costs they must pay, so miners have extra incentive to follow the money.

Keep in mind that segwit support amongst miners languished for months around 30% before shooting up practically overnight. While miners may signal for one thing, when push comes to shove they historically have always avoided being on the wrong side of a chain split:

Segwit activaition looked very unlikely to meet the required 80% threshold for a long time (source)

Developers

If you check the commits on the btc1 repo, you can see that it is not very actively developed. Compared to the bustling bitcoin core repo, it is a ghost town.

A common argument from B2X supporters is that the (largely unpaid) contributors to Bitcoin Core will start contributing to B2X after the fork. That’s pure fantasy:

Click the tweet to see the overwhelming statements backing this up

B2X is betting that it will attract new talent, but I wouldn’t hold my breath. It’s a very complicated protocol with a steep learning curve.

It’s a bold move to switch from a cryptocurrency with an experienced team and long track record of stable software, to a new one with almost no developers.

Negative Externalities

After how well things have gone since the Bitcoin Cash fork, it’s easy to believe that future hardforks will be equally successful. Keep in mind that divided hardforks cause a number of problems:

• Naming: with bitcoin, bitcoin cash, B2X, and (soon) bitcoin gold, it becomes hard for newcomers to understand what’s going on.
• User Experience: not only do these blockchains share the same transaction history, but they also have the same address format, which can lead users to send funds on the wrong blockchain!
• Taxes: it is still unclear whether regulatory bodies will treat these as airdrops (potentially taxable) or stock splits (not taxable).
• Privacy: since multiple blockchains now share the same transaction history, future transactions on one blockchain can leak privacy info on another. This is very hard to manage.
• Company Support: many services are effectively forced into supporting alt-coins that they otherwise would not, which includes extra engineering, accounting, security, and support work. It may not even be a choice for these services, as one could argue they have a fiduciary duty to pass along the fruits of holding onto their end-users’ private keys.
• Security: having a large supply of performant SHA256 ASICs in the world makes it easier to launch a 51% attack on bitcoin. As a side note, bitcoin cash’s poorly implemented Emergency Difficulty Adjustment (EDA) leads to unnecessary swings of hashpower between blockchains, but that mostly just hurts Bitcoin Cash.
• Blockchain Bloat: uncertainty about how to prevent replay transactions may lead to users splitting their coins in transactions between their own wallets before spending their funds, polluting one (or both) blockchains a bit.

Last time, there was an enormous difference in beliefs between people who wanted to scale with much bigger blocks (bitcoin cash) vs segwit (bitcoin core). This time, the two factions are arguing about a relatively small difference (1MB vs 2MB base blocks), and this is taking place after we just effectively had a doubling of the effective block capacity (via segwit). The debate is really how we think Bitcoin should change, and who we think has the power to change it.

Is this good or bad?

Yikes

So, what’s going to happen?

I sure wish I could predict the future! I don’t know for sure, but here are my best guesses…

There Will Be Downtime

Bitcoin services will voluntarily stop transacting on the network before the hardfork, and maintain silence for some time afterwards while they wait for the dust to settle. There may be large trading opportunities that are only available to users with pre-funded balances held at various custodial services.

Bitcoin Core (Legacy) Chain Won’t Die

In the unlikely worst case that the bitcoin blockchain were actually reduced to 15% hash power, it would be much harder to use. ~10 minute blocks would now take ~67 minutes, which would restrict transaction throughput substantially and cause transaction fees to go up. The difficulty adjustment that normally takes two weeks (2016 blocks) would now take 13+ weeks, so these slow block times (and low throughput) would persist for a while. This is is not good. While the user experience for transacting would be awful for months, HODLers would be unaffected (the demographic that supports legacy bitcoin). Keep in mind that in order to actually mine B2X you have to install new software (vs just continuing to run your old software to mine Bitcoin).

The above assumes that miners do not care about their own profits. If prediction markets can be trusted, after the fork Bitcoin Core will be worth ~6x as much money. Even a strong cartel would be tempted to run some (most?) of its hashing power anonymously on the legacy Bitcoin network to get some of that sweet sweet profit. After a painfully slow 2,016 blocks, legacy bitcoin’s difficult would drop (assuming no change in hashpower), making legacy bitcoin mining an even better deal (assuming no change in price).

Bitcoin Supporters Have Always Been Very Passionate and Will Remain So

There are a lot of companies that don’t like B2X and bitcoin core has a fanatical user base, many of whom run their own full nodes.

However, if the futures markets were wrong and legacy bitcoin is worth a lot less than B2X, that would dramatically change the landscape. It could become so unprofitable to mine bitcoin that the coin stalls out as it is too expensive to mine a single block. At that point, the only way legacy bitcoin could survive would be a proof of work change (aka the nuclear option):

This is the potential problem bitcoin gold is trying to solve. Premature optimization or useful insurance policy?

B2X Chain Seems Likely to Launch

B2X has support from some of the biggest companies in the space (whose businesses would be better served by cheap on-chain transactions): BitPay, blockchain.info, Coinbase, Xapo, ShapeShift, and more (full list here). B2X also has the support of ~85% of miners.

This is the culmination of years of debate, and there’s a lot of pride and ego at stake. It has become a somewhat emotional issue. Simple conversations quickly escalate into childish name-calling:

This is the CEO of a company that may be worth $1 BN! Who would’ve thought a discussion on blockchain scaling would be NSFW?

It Is Possible the B2X Chain Dies Quickly

While unlikely, B2X supporters are quite mute given the number of signatories to the New York Agreement. Many of the people who strongly prefer bigger blocks left the ecosystem to work on other projects like bitcoin cash and altcoins long before that.

Beyond Jeff Garzik, it’s hard to find prominent bitcoiners actively advocating for B2X today. Xapo and Coinbase are prominent NYA signatories and each put out statements (see Xapo here and Coinbase here) on how they’ll support both forks for users, but neither has actively voiced support for B2X recently. I was surprised that some NYA participants who spoke to me off the record did not want to give statements on the record (keep in mind that it’s a very divided community and social media has unfortunately become a bit toxic). If any of them would like to, I will happily publish them here.

BitGo CEO Mike Belshe who just published piece in support of Segwit2x called “Keeping the Community Together”, though the company seems to be in disagreement:

Is this what your water cooler conversations are like at the office?

Roger Ver (aka “Bitcoin Jesus”) agreed to trade 1,000 BTC (!) for 1,000 B2X spread across smaller whales Ben Davenport, Charlie Lee, Alex Morcos, and Tuur Demeester:

Kids growing up these days are going to be very confused when they learn that gambling is in fact illegal.

Trace Meyer has offered up a 25,000 BTC (!) trade, which Roger has not responded to:

That’s like a gazillion satoshis!

B2X is betting on having majority hashing power and that SPV nodes will follow. If majority hashing power doesn’t pan out and the value of the B2X coin is as low as futures markets predict, it may not be worth it to miners to expend resources on hashing. Without an emergency difficulty adjustment, B2X could be DOA.

Replay Protection is Going to be a Problem

While bitcoin services should be smart enough to protect themselves against transaction replay, some users who manage their own private keys are going to inadvertently have their transactions replayed. Expect to see lots of horror stories about losing money popping up on reddit, twitter, and elsewhere. Hopefully, most companies will (eventually) refund most of these transactions, but P2P users may not be so generous.

Unfortunately, the winner here is going to be altcoins :(

When the Dust Settles, the Blame Begins

After the fork has played out, the best metric for success will be price.

If the new Segwit2x coin is worth more than legacy bitcoin, it will have won. The old coin will still be used by some believers, while the new segwit2x coin may be crowned as the one true bitcoin. There will be lots of bitterness between the camps, but the good news is that they can go their separate ways. The Segwit2x camp will be able to claim victory, and defend why the lack of replay protection was needed in the first place. Those against the majority hardfork will be labelled as enemies of bitcoin and will be vilified for getting in the way of progress.

If the new Segwit2x coin is worth less than legacy bitcoin, it will have lost. The new coin may still be used by some believers, or die out altogether. There will be lots of bitterness between the camps, and core will have more soft power in guiding the direction of the protocol going forward. We may see a rapid release of new features (aggregated schnorr signatures, merklized abstract syntax trees, sidechains, IBLT, weak blocks, and more). Proponents of the minority hardfork will be labelled as enemies of bitcoin and will be vilified for getting in the way of progress.

Honey badger don’t care… about which fork you favor.

Get Me Some Popcorn

This is going to be quite a show. There will be arguments over which is the real bitcoin, the price of one/both forks could crash or rally, block times will decrease, miners could jump ship (and lie about it!), rushed new software could have bugs, and much more.

The 21st century equivalent of a soldier’s war drum. Are you intimidated?

Disclaimer

Keep in mind that this situation is in constant flux. It’s amazing that we’re less than a month away from the biggest change the bitcoin network has ever seen and still don’t know what (if any) replay protection will exist on B2X. Bitcoin Core has always been very slow and deliberate in making changes, with a strong bias towards backwards compatibility, while B2X is taking a different approach.

We’re in uncharted territory, and with a currency that is currently worth almost one-hundred billion dollars.

That’s a lot of money! Only a cartoon madman would be cavalier about that.

Thanks to everyone who reviewed and provided commentary on this post.