Removing Bitcoin’s Guardrails
Abstract: We look at the recent controversy related to the potential removal of the OP_Return policy limit in Bitcoin Core. We reflect on the economic reality, that miners want to produce the most profitable blocks, regardless of whether the blocks contain spam or not. We note that if a spammer wants to outbid other users, they can. As evidence for this, we note that miners have received over 7,000 bitcoin in fees related to Ordinals. We argue that those running nodes may want an effective mempool, one that more reliably predicts what miners may mine, to help users receive new blocks faster and determine appropriate fees for their transactions. We are therefore somewhat supportive of removing the limit.
The OP_Return Limit
In the Bitcoin Core software repository, it has recently been proposed that the limit on the size of OP_Return outputs should be removed. This is an output type designed to store arbitrary (non-transaction) data in Bitcoin’s blockchain. With the limit in place, Bitcoin Core nodes would not relay transactions which breached this limit. However, it is not a consensus rule and Bitcoin Core nodes would always accept a block containing such a transaction output, as a valid block.
The removal of this limit proved somewhat controversial, with some arguing that the limit is necessary to deter or prevent spam. Many advocates of keeping the limit in place, want Bitcoin to be used only for “financial transactions” and want to keep things like image related data as non-standard transactions.
The Ordinals Boom
However, there has already been a boom in storing images on Bitcoin’s blockchain. This boom happened at the start of 2023 and was called “Ordinals”. Instead of using OP_Return outputs, the images are stored in the input script for Taproot spends. The majority of these Ordinals transactions are already standard, they are relayed by Bitcoin Core nodes. In some circumstances this Taproot method is cheaper than OP_Return, because they benefit from the 75% witness discount from the SegWit upgrade.
Ordinals usage has been strong, according to a dashboard on Dune.com, there have been over 88 million inscriptions, paying over 7,000 bitcoin in transaction fees. Based on the current Bitcoin price, this is over $660 million. Businesses have been set up in the Ordinals space, hoping to capitalise on the growth and millions have been invested into Ordinals related tools, such as wallets, inscription trading systems and methods of creating Ordinals and submitting them to the network.
Many regard these images as spam. In our view, in this context, we like to think about spam from the perspective of the intent of the person creating the transactions. Are they trying to deliberately cause harm to others or are they trying to benefit personally? With this in mind, we do not think storing images in the blockchain is always spam, as the people doing this seem to mostly be doing it for their personal enjoyment or to speculate and to try to profit. However, to the extent that people are putting images on the blockchain to troll others, which is happening to some extent, then yes it is spam.
While images on the blockchain may seem spammy to many, we agree with the Subjective Theory of Value:
The value of various consumer goods and services does not reside objectively and intrinsically in the things themselves, apart from the individual who is making an evaluation. His valuation is a subjective matter that even he cannot reduce to objective terms or measurement.
Some people seem to like images on the blockchain and have spent over $600 million in fees on them. Who are we to question that, when the value of these goods is subjective? All we can say is that we do not value these images and that we won’t be paying for them. In our view, it’s highly likely that businesses and people investing in this area with the hopes of profiting, will end up losing their money. But let the market decide!
Our point is that the horse has bolted, people already do use the blockchain for images on a mass scale and keeping the OP_Return limit won’t change that. The systems are already in place to use part of the Taproot input script for images, which is 4x cheaper per byte than OP_Return anyway.
Bitcoin Mining
We have been following the Bitcoin mining space for over 14 years. We have seen Bitcoin mining transition from a hobbyist endeavour to an industry of listed companies. There are 10 of these public companies we follow closely and we read pretty much all of their public disclosure documents. These miners all report their financial statements each quarter and update the market each month on their production figures. We regularly speak to the investment relations officials and the management teams of these companies. These management teams love Ordinals, they see them as a potential revenue driver. A critical revenue driver in a tough competitive industry. The idea that Ordinals are spam and should be filtered, does not and will not resonate with these professional management teams. Some may not like it, but this is the business reality. This is the reality some of us always expected. Bitcoin has grown, it’s a business and businesses need to maximise earnings and return on equity.
Bitcoin is about incentives and aiming for incentive compatibility. Bitcoin doesn’t work because the space consists of a group of altruistic and well meaning people, with common goals. The Bitcoin space consists of a wide diversity of people, with different perspectives and philosophies. Bitcoin doesn’t work because we are all on the same team, it works because it is robust and incentives are aligned. In our view, it’s time to remove the paternalistic guardrail of the OP_Return limit and embrace the economic reality of blockspace markets.
If larger OP_Return outputs are kept non-standard but people still want to use them anyway, miners will just launch businesses which receive these transactions directly, bypassing the public memory pool. The largest listed miner, Marathon [MARA US] has already done this. However, our understanding is the service Marathon provides is currently not popular. Nevertheless, if miners start to receive transactions out of band, this has many negative consequences for Bitcoin. This would mean that the differences between the transactions in the blocks miners produce and what users expect to see, will increase. This could break technologies like Compact blocks, which helps blocks propagate across the network faster, by removing the need for nodes to download transactions twice (once for the mempool and again once it’s in a block). It is probably sensible for Bitcoin Core to preemptively remove the limit to make sure Compact blocks do not break. If it does break and block propagation delays increase, then this could benefit larger miners and larger pools, at the expense of smaller miners, increasing mining centralisation.
Miners and pools launching businesses to accept non-standard transactions also has other negative consequences. There are costs in setting up such businesses, technical costs and marketing costs for instance. This business model is also potentially monopolistic, with users wanting to use one simple platform to submit their non-standard transactions. This increases the barrier to entry in mining and mining pools, making mining more difficult for smaller players. Which again, causes more centralisation pressure. Once these systems gain traction, it could be challenging to stop these businesses, even if Bitcoin Core policies are later relaxed. For example once the infrastructure is built, lazy customers could continue using these services instead of the public mempool.
In our view, it is desirable for Bitcoin developers to try and keep the software competitive. To make the open source transaction selection algorithm competitive at maximising revenue, to stop miners building their own proprietary algorithms and also to make the public mempool competitive, to stop miners building private mempool businesses. We appreciate that not everyone sees it that way, but this is the economic reality of the mining space now. We want the mempool to work effectively and in our view removing the OP_Return limit is the better option, rather than putting one’s head in the sand and trying to pretend that spam transactions are not getting mined. The choice is to have an effective mempool or an ineffective one.
Node Runners
If one assumes that the blockchain is full, then increased usage of OP_Return actually makes it easier to run a full node. Remember, OP_Return doesn’t benefit from the witness discount, therefore the maximum size of a block consisting of OP_Return outputs is 1MB, far smaller than the 4MB maximum. At the same time, the OP_Return outputs do not bloat the UTXO set. Other protocols use alternative systems such as fake addresses to store arbitrary data and this methodology has serious negative consequences for those seeking to validate all of the Bitcoin transactions. OP_Return is just data that does not require verification and can then be disregarded. Those concerned about cheaper node operating costs have nothing to worry about from the removal of the OP_Return limit.
How do we stop spam?
We will start this section by quoting from Eric Voskuil’s book “Cryptoeconomics”. In the book Eric says that:
Resistance to censorship is a consequence of transaction fees.
A core objective of Bitcoin is censorship resistance and transaction fees are a fundamental part of the security model that aims to achieve that objective. Adversaries wishing to censor transactions cannot succeed by encouraging node runners to filter out certain transactions from their memory pools. If that could work, Bitcoin would not be especially good at preventing censorship. Instead, it is all about fees, feerates and miners trying to maximise their revenue per block. This applies whether the transaction is spam or not, whether the transaction uses OP_Return or not or whether the transaction uses a segregated witness or not. To get into the blockchain, one needs to outbid other users, this is the only viable spam prevention model. Or as Satoshi put it:
At some price, you can pretty much always get in if you’re willing to outbid the other customers
This spam prevention model has been understood for years. As we said in our September 2017 piece on SegWit:
It is true that a spammer could produce such a 4MB block, which would cost the same in fees as a 1MB block. This is a potential problem. However, this is not a change to the security properties of the system, because the 4MB block is not cheaper than the 1MB block, it merely costs the same. A spammer can always outbid legitimate users, regardless of SegWit. Indeed an attacker could simply produce 1MB of non-witness data, to compete with “legitimate” users, and this attack vector is no cheaper than before. SegWit does not and cannot change the security dynamic, that if an attacker wants to outbid users with spam data, they can.
To make sure your transaction gets confirmed in a timely manner, you need good information about what miners may mine in the next block, so that you can set the appropriate transaction fee. This is another reason that node operators may want their mempool policy to match what miners are actually doing as closely as possible. If Bitcoin Core doesn’t remove the OP_Return relay limit, users would have to run other software or use third party websites to get information on what transaction fees to use.
It is true that the OP_Return relay limit was put in place in 2014 and Bitcoin has worked fine for 11 years with this limit. So why remove it now? In our view this reasoning feels a bit like a discussion that was often had during the blocksize war in the 2015 to 2017 period. “Large blockers” often said that Bitcoin had worked fine for years without full blocks, so why introduce full blocks now? The answer is that this is just the economic reality of success. As former Bitcoin developer Gregory Maxwell said in an important email in 2015:
Demand for cheap highly-replicated perpetual storage is unbounded
In our view, these changes were always inevitable. There is unlimited demand to store images in the blockchain and the only way to deter that is with transaction fees. It is just that some people always saw it that way and others did not, which can lead to arguments.
We are fortunate that the small blockers won the blocksize war. If the large blockers had won and say Bitcoin XT was adopted, the blocksize limit would now be around 250MB. The blocks would potentially be full of images, 250MB of images every 10 minutes. This may have made running a node unfeasible for ordinary users and may have killed Bitcoin by now.
Conclusion
We are somewhat supportive of the removal of the OP_Return limit. It is time to face up to the economic realities and be competitive. We want local mempools to be effective and for the public p2p transaction broadcast system to be the winner. If an attacker or spammer wants to outbid other users, they can and we should embrace that reality. Spam budgets do not last forever and many of the people investing in blockchain images are likely to collectively lose millions of dollars. People will learn hard lessons and then Bitcoin will be stronger from it.
The post Removing Bitcoin’s Guardrails appeared first on BitMEX Blog.
BitMEX Blog
