Why I don’t like artificial block space scarcity

In recent months, there has been a great deal of debate about the fee market, block capacity, and growth on the bitcoin network. I’ve previously written about that debate.

Today, I want to focus on the ‘fee market’, or what think should be more accurately called artificial block space scarcity.

On one side, some feel that since a fee market may need to develop eventually, we should let it develop now. In essence, Bitcoin-Core has already embarked on this plan and we are beginning to see the realities of it in practice.

On the other hand, others feel that there is little need for a fee market now and prematurely attempting to develop one will lead to constrained growth. Much of the debate on both sides has been fairly theoretical and academic. Theory is great, but in the real world, the practical consequences of applied theory aren’t always as successful as they were predicted to be on paper.

In this post, I’m going to focus on practical realities of the nascent fee market that has been developing over the last few months and how emerging trends indicate that the fee market, as it stands now, will most likely make not just Bitcoin less useful for users, but will lead to a less open bitcoin network. It’s those practical realities that make me dislike the current scarcity, or ‘develop a fee market now’ approach.

So why do I dislike artificial block space scarcity?

  1. It makes sustaining spam attacks easier

If a glass is 80% full, its much easier to fill it the rest of the way than if it was 20% full. While we can debate the exact math, it’s a certainty that less used space makes attacks cheaper and thus easier to continue over a long period. These attacks have a real cost on the network for node operations, miners, and anyone trying to use the network for normal business or personal activity.

2. ‘Dynamic Fees’ are not the solution

I’ve often heard the suggestion that the solution to block scarcity is easy — just pull in dynamic fees from Bitcoin-Core! Unfortunately, this is at best a half measure. Dynamic fees currently look only backwards, which is a weak way to predict at best.

It is possible to improve fee prediction, as compared to the fee prediction functionality in Bitcoin-Core .0.12, but unfortunately, most of the data you need to do this (transaction backlog, send rate, etc) is held by a few companies, ours among them, and isn’t publicly available on the network. Setting this and many other problems aside, the most significant issue is that the fee market isn’t a real time market or open auction at all.

3. The fee market is not a real time market

Predictability is perhaps one of the most important things for a business serving customers. Customers expect a reliable experience. If you are a large bitcoin company, there is an alternative to bidding every few seconds for space with a dynamic fee — enter into a priority contract with a large miner. In short, a company prepays for block space and the miner, or mining pool, guarantees that regardless of market rate for transactions at the time, it will confirm the companies transactions first.

This is entirely rational behavior — the company receives predictability for its products and users and the miner receives more predictable revenue. As the network gets more congested, we will see an increasing number of these deals. Indeed, two such arrangements have already been publicly announced by Huobi and BTCC (see links at bottom). I am reasonably certain more such deals have already been made in private and more will be. Indeed, this will likely be a race to zero in which companies either enter similar agreements or let their products suffer. As time goes on, most space will be bought in advance, which will make even predicting fees next too impossible, as it will be hard to tell which prices are in fact part of the real time market.

4. Block space scarcity is bad for the openness of the network

The simple fact is, larger startups can most likely engineer or contract around these issues. It won’t be easy, and new product development may slow down, but the pain is more likely to be felt by smaller startups and folks running their own full node wallets like Electrum. Historically, the barrier of entry onto the bitcoin of network has been very low, but the premature development of a fee market is likely to raise the barrier of entry by several orders of magnitude. Small projects and individuals will be priced out.

I’ve previously written about how I think bitcoin will offer less value to users as the network gets congested, but the above worries me even more. The prospect of a less open, less innovative, and more centralized bitcoin is deeply troubling.