The Big Blockers
The current state of the scaling debate presents itself as follows. Bitcoin’s transaction throughput is roughly (2000 transactions / MB) * (1 MB / block) * (1 block / 10 min) = 3.3 transactions / sec. This throughput became a bottleneck lately and led to high fees and long confirmation times. There are two strategies to increase this throughput:
Increase the “2000 transactions / MB”-factor by means of various optimizations and second layers, such as SegWit and Lightning Network.
Increase the “1MB / block”-factor by means of a backward-incompatible protocol change
Two camps formed around these strategies, the SmallBlockers and the BigBlockers, fighting each other fiercely in social media. An outsider inevitably wonders why this seeming question of taste is so important to these camps that each is rather willing to split Bitcoin than to give in. In order to answer this question one needs to know that the Big-Block-Strategy ..
.. allows to increase the (on-chain) throughput stronger and in a shorter time than the Small-Block-Strategy
.. furthers centralization (running a full node becomes more expensive, hence full nodes become less)
.. will split Bitcoin into two coins (increasing the block size requires a hard fork; the mere existence of SmallBlockers makes this hard fork contentious; every contentious hard fork results in a permanent chain split)
Most Bitcoiners consider the Throughput-Property as positive and the Centralization- and Split-Property as negative. I will call these Bitcoiners good. Those who assess the above properties differently will be called malicious. With this, BigBlockers can be categorized into three classes:
- Good BigBlockers with scaling intention
These are good BigBlockers whose decision to become a BigBlocker is based on the trade-off between the above properties and nothing else. They must necessarily be of the opinion that the benefit of the Throughput-Property trumps the drawbacks of the Centralization- and Split-Property.
- Good BigBlockers with hidden motives
These are good BigBlockers who wouldn’t promote the Big-Block-Strategy if it weren’t for reasons other than the Throughput-Property. Despite being “good” (by definition), they are, of course, bad actors.
- Malicious BigBlockers
Malicious BigBlockers have agendas such as controlling or destroying Bitcoin. Because their agenda has no majority appeal among Bitcoiners, they usually pretend to be good BigBlockers with scaling intention. (One notable exception is Craig Steven Wright. While dishonest in every other way, he makes no secret of approving centralization.)
In the following I will argue that some prominent BigBlockers are not good BigBlockers with scaling intention and I will make some educated guesses regarding their true motives.
There have been long-running spam attacks on the Bitcoin Network that remained undetected up until recently. These spam attacks consumed a considerable amount of transaction throughput. It is not clear whether these spam attacks were motivated by the scaling debate. However, if they were, then the spammers belonged to the BigBlocker camp: The consumed throughput tightened the urgency for scaling and, this way, helped the BigBlockers by providing more weight for the Throughput-Property. It is a remarkable coincidence that the maximum transaction backlog happened at about the time of the New York Agreement. Since the spammers intentionally produced congestion to push their agenda, they are obviously not motivated by the goal to relieve the network from this very congestion and therefore are malicious BigBlockers.
Decent spam attacks may well consume several hundred BTC in transaction fees; only a financially strong player can afford them. Interestingly enough, miners are able to spam at a discount because they retrieve their spent fees in part. A spam attack may even become profitable for a miner if he owns enough hash power. As an example, consider a miner with 40% hash power who conducts a massive spam attack that fills 25% of each mined block with its transactions. Imagine further that the resulting squeeze of block space pushes the fees per block from 1 BTC to 3 BTC on average. Before his attack, this miner collected on average 0.4 BTC in fees per Block. Afterwards, he collects 1.2 BTC in fees per Block and pays 0.75 BTC per Block for his attack. Thus, the spam attack earns him a net profit of 0.05 BTC per block.
Craig Steven Wright / nChain
Craig Steven Wright claims to be Satoshi Nakamato, the inventor of Bitcoin. He presented a proof for his claim, but this proof turned out to be a scam. Since then, he went along the lines “I still can prove it, but I don’t want to”. Before his fake coming out, he managed to trick former Bitcoin Core lead maintainer Gavin Andresen into believing him and to publicly endorse him. This endorsement temporarily earned Wright some credibility, and permanently damaged Andresen’s. Quite fittingly, Wright is employed with nChain, a company with bold claims but nothing to show.
Wright doesn’t consider the Centralization-Property as negative and therefore belongs to the malicious BigBlockers. To quote his own words: “That’s what Bitcoin is about: Hard, central, controlled, no-one-can-change money”. Later on in the same speech he clarifies: “If you [..] can’t afford a $20.000 node to help this network, piss off.”
Now, you would think that hardly anybody takes Wright serious and it’s not worth mentioning him. However, curiously enough, there are players in the BigBlocker camp who still assign value to keeping Wright in the game: While there is no direct evidence that Wright/nChain is affiliated with Bitmain, it is remarkable that Bitmain gives Wright a stage (the above quotes stem from a Bitmain sponsored BigBlocker conference) and Wright praises Bitmain’s “Bitcoin ABC”.
Jihan Wu / Bitmain
Jihan Wu is the face of Bitmain and one of its two CEOs. Bitmain produces mining hardware (AntMiner) and managed to crowd out most of its competitors. It also operates mining pools (AntPool, BTC.com, ConnectBTC) which, at the time this was written, account for one third of the total hashrate. If BTC.TOP and ViaBTC are also counted towards Bitmain, then Bitmain controls the majority of hashpower. How did Bitmain become so successful? The CEO of one of Bitmain’s former competitors said in January 2017: “We have tried to calculate the amount of money that the Chinese have invested in mining, we estimate it to be in the hundreds of millions of dollars. Even with free electricity we cannot see how they will ever get this money back. Either they don’t know what they are doing, but that is not very likely at this scale, or they have some secret advantage that we don’t know about.”
I suggest two strong candidates for what this secret advantage might be:
- Theory 1: Covert ASICBOOST
Covert ASICBOOST is a method to secretly save mining work. Bitmain admits that it holds a patent for ASICBOOST, that it implemented ASICBOOST in its chips, that it used ASICBOOST in testnet, and yet claims that, “for the greater good of Bitcoin”, it would never use it in mainnet. It is hard to believe that a company puts so much effort in a project without having the intention to deploy it. It also doesn’t fit to Bitmain’s opposition against an inhibition proposal for covert ASICBOOST on the ground that “this would be a loss for the patent owners”. Finally, one of the telling signs for the covert use of ASICBOOST is a high number of (near) empty blocks and, surprise, surprise, AntPool mines a lot of these.
- Theory 2: Governmental Subsidies
China subsidizes selected companies in order to transform them into national champions. National champions are companies which are supported by the government and, in return, act in line with the government’s strategic aims, even at the expense of profits. The governmental support allows national champions to become dominant players, domestically and even globally. It naturally comes into mind that Bitmain might be a national champion.
Jihan Wu doesn’t care much about Bitcoin’s transaction throughput. While he was crying for bigger blocks, his AntPool mined one empty block after another with no intention to improve this behavior. (Bitfury, on the contrary, although it never complained about the block size, voluntarily refrained from mining empty blocks in order to maximize the transaction throughput.) Despite acknowledging that SegWit would alleviate the “block size scarcity problems”, Wu stalled SegWit for more than 8 months in order to pressure for bigger blocks (until the BIP 148 UASF forced him to give in). So, obviously, Wu’s appetite for bigger blocks doesn’t originate from the Throughput-Property and, for that reason, he is not a good BigBlocker with scaling intention.
But then, what is Wu’s motive for the Big-Block-Strategy? Bitcoin Core developer Gregory Maxwell is positive about Theory 1 and pointed out that Bitmain might gain up to $100 million USD per year from using covert ASICBOOST. Since SegWit is incompatible with covert ASICBOOST, there are a hundred million reasons for Bitmain to block SegWit. If Maxwell is right, then Wu did not stall SegWit in order to pressure for bigger blocks, but rather he insisted on bigger blocks in order to justify stalling SegWit. However, there are calculations which suggest that Bitmain’s gains from covert ASICBOOST are much lower. Also, from the fact that Wu signed the 2016 Hong Kong Agreement, it appears that Wu is indeed willing to trade SegWit for bigger blocks and, therefore, his interest in bigger blocks seems to be genuine. So, the question remains: Why is Wu interested in bigger blocks?
If Theory 2 is correct, then Bitmain’s politics is largely dictated by the Chinese government. The recent intervention of the PBoC, which led to a months-long withdrawal freeze on major Chinese Bitcoin exchanges, clearly shows that China views Bitcoin with suspicion. Without doubt, the Chinese government considers the Centralization-Property as positive, because it paves the way for controlling Bitcoin, and it considers the Split-Property as positive, because it weakens Bitcoin. Thus, the Chinese government is a malicious BigBlocker and if it acts upon Bitmain, then Wu’s behavior suddenly makes sense.
SmallBlockers practicing game theory in the context of the scaling debate should keep in mind that a “nuclear escalation” might be an acceptable, if not preferred scenario for a government-controlled Bitmain. From this perspective, also the Antbleed backdoor appears in a new light.
Roger Ver / Bitcoin.com
Roger Ver bought several ten thousand Bitcoin when Bitcoin’s price was at around $1. Later on he invested in various Bitcoin related startups and became known for his passionate advocacy for Bitcoin. About two years ago, he became an equally passionate advocate for the Big-Block-Strategy. So, what kind of BigBlocker is Ver?
Ver is not a technical guy. He doesn’t understand the mechanisms and implications of soft and hard forks. (For example, in an online discussion last December he spoke of an “economic hard fork”, worried over SegWit being “a soft fork that can’t be undone” and wasn’t aware of the fundamental difference between the SegWit soft fork and the hard-forking Bitcoin XT, Bitcoin Classic and Bitcoin Unlimited, which he actively supported.) Due to his lack of understanding, Ver is utterly unafraid of hard forks. Therefore, Ver’s stance regarding the Split-Property is probably best described as neutral.
Ver denies the Centralization-Property. He is of the opinion that bigger blocks further decentralization, and not centralization. His reasoning behind this claim is that bigger blocks would attract more users, and more users would spawn more full nodes. Since at least he doesn’t question the merit of decentralization, one might assume that, if he acknowledged the Centralization-Property, he would consider this property as negative.
Ver complains day in and day out that the fees are too high and the confirmations take too long. His focus is on an immediate transaction throughput increase. So, obviously, he considers the Throughput-Property as positive. In fact, the Throughput-Property is apparently the only property that is relevant to him.
According to how Ver assesses the properties of the Big-Block-Strategy, he is, by definition, a good BigBlocker (ignoring his neutral stance on the Split-Property). Now, the interesting question is whether Ver’s commitment for bigger blocks is based only on the consideration of these properties or whether he is influenced by additional incentives.
As mentioned above, Ver is invested in several Bitcoin related startups, such as BitPay and Blockchain. Unlike Bitcoin itself, these startups are heavily reliant on smooth transactions and a fast Bitcoin adoption and, for this reason, are vulnerable to high fees and long confirmation times. Chances are that, when weighting the Throughput-Property, Ver hasn’t only the good of Bitcoin in mind but also (and maybe primarily) the good of his companies. Nevertheless, it could still be argued that these companies served Bitcoin well and that their continued existence is in Bitcoin’s best interest. Clear evidence that Ver is not acting in Bitcoin’s best interest can be found by looking at his stance towards SegWit. While Bitcoin Core developed SegWit, Ver supported several attempts (Bitcoin XT, Bitcoin Classic, Bitcoin Unlimited) to replace Core, because Core didn’t meet his desire for bigger blocks. These attempts were to no avail and last November SegWit was finally ready to take off. Among other benefits, SegWit allows for a near-term throughput increase of approximately a factor of two. Remarkably, instead of embracing this throughput increase when it became available, Ver sabotaged SegWit in order to retaliate against Core for having denied his bigger-blocks-request.
So, all in all, Ver’s fight for bigger blocks is not only based on the merit of the Throughput-Property but, crucially, on deficient technical knowledge, business interests, and hostility towards Bitcoin Core. For this reason, Ver is a good BigBlocker with hidden motives.
Some prominent BigBlockers are not good BigBlockers with scaling intention. Those of them who pretend to be good BigBlockers with scaling intention obviously have a hidden agenda. Possible hidden motives were discussed, but, ultimately, this can only be speculation.
If you consider yourself a BigBlocker, then be aware that some of your most vocal fellow campaigners don’t act in good faith. If you don’t have a strong opinion about the block size and wonder why the scaling debate continued for so long and was fought so fiercely, so uncompromisingly, then here’s your answer: It never was about scaling.