Despite a record number of transactions, the Bitcoin blockchain works much better than during the last big bull run of 2017
The number of daily Bitcoin payments has recently reached a new All-Time-High, while the MemPool of unconfirmed transactions remains surprisingly small. The reason for this is the increasingly efficient use of space on the blockchain through a method called batching.
Despite its ever-stronger status as digital gold, Bitcoin remains first and foremost a payment system. And a payment system can be judged by how many payments it processes. It fits well with the All-Time-High of the Bitcoin price that the number of payments has reached a new All-Time-High, too.
However, you have to search for a while to find this in the charts. If you look at the number of daily transactions, for example on blockchain.com, you might be disappointed: There was a small spike on January 7 with 401,000 daily transactions, but the value is still far away from previous records, such as May 2, 2019 with more than 450,000 transactions.
Anatomy of a Bitcoin Transaction
So why has Bitcoin reached a new record for daily payments? To understand this, you need to understand what a Bitcoin transaction is like. Bank transfers have a sender and a receiver, which technically could be called an input and an output.
Bitcoin transactions, on the other hand, can have any number of inputs and outputs. This is called a UTXO system: when you form a transaction, you take a certain number of coins, melt them down and use them to form the outgoing coins. Metaphoracally, there is a mint press in every transaction.
A single transaction can have multiple outputs, and thus multiple recipients. Since it doesn’t make much sense and is also very impractical to accept a payment in the form of multiple coins, we can assume that each output of a transaction represents one payment. One therefore recognizes the number of daily payments not by the number of transactions, but by the number of outputs. This is exactly what Transactionfee.info does to plot a chart of daily payments. And this chart reached a new all-time high on January 9 with nearly 800,000 payments.
More Payments per Transaction
Bitcoin is processing more payments while the transaction volume remains constant, as transactions contain more outputs. This can also be demonstrated statistically. At the beginning of 2020, a transaction contained an average of 1.5 outputs. Today, the figure is around 2.2, an increase of almost 50 percent.
The main driver of this development is the so-called batching by platforms such as exchanges. Instead of processing each payout immediately, they collect them to accommodate a whole batch of payments in a single transaction. This, as we will see later, saves significantly more fees than one might suspect at first glance.
Bitcoin is scaling inward rather than outward: instead of taking up more space, it is using space more efficiently. You can compare it to putting more seats in a bus instead of building a bigger bus, or to compressing files instead of adding a new hard drive.
A Relaxed MemPool
All of this works surprisingly well. The MemPool of unconfirmed transactions remains calm despite the hype and the bubble. At the end of 2017, when we had the last big bubble, this looked quite different. At that time, the rise in price was accompanied by a dramatic rise in fees. Sometime it would cost more than 50 USD to send a simple transaction. Such a fee bubble can currently be observed on Ethereum, where the price of simple transactions can be more than 10 USD, while the execution of smart contracts costs 100 USD or more.
For Bitcoin the situation is much more relaxed, given the circumstances. The blockchain is indeed becoming fuller, the size of blocks has reached an all-time high of 1.3 MB on average. But those who are not in too much of a hurry can get by with a fee of around 90 cents; those who absolutely want their transaction in the next block have good prospects with a fee of around 5 USD. Even in the busiest moments of the past few weeks, it never got much more expensive. Those who postpone their payment until the next week-end can even get away with a fee of 20 to 30 cents.
At the same time, the volume sent per transaction has become significantly higher, as statistics from Bitinfocharts.com show: While the average volume per transaction was just over 100,000 USD in December 2017, it is now more than 200,000 USD (another All-Time-High). This means that the fees are still quite moderate in relation to the amounts being sent.
This assumes that one sends “simple” transactions that issue a single input. However, transactions which require multiple inputs quickly multiply the amount of fees. Let me explain why.
Inputs Cost More Than Outputs
As mentioned, an input is a “coin.” For example, if you send 0.001 Bitcoin (about 36 USD) to my Bitcoin address 1BergmanNpFqZwALMRe8G-HJqGhtEFD3xMw, the amount will be in my wallet as a 36 dollar coin. If someone sends another 0.0005 bitcoin to the same address later, it will be in my wallet as a second coin.
If I subsequently pay for something with Bitcoin, I can spend a single coin for amounts up to 36 dollars. One input, low fees. However, if I pay more than 36 dollars, I have to merge the two coins. Two inputs, higher fees.
Each input requires a signature, as the owner of the coin must prove that he or she owns it. Therefore, the transaction must be signed with the key associated with that coin. These signatures make up a large part of the data in a transaction. In Bitcoin the fee of a transaction depends on its size in bytes – not on the amount of money being sent, as one might expect. An output is just the information about the next owner of that coin, and is therefore much cheaper.
Here are some examples: A standard transaction with one input and two outputs is 220 bytes in size. If you increase the number of outputs to six, the size only increases to 344 bytes. If the transaction contains six inputs and two outputs, it is 960 bytes in size. More outputs increase the transaction size only slightly, while more inputs increase it enormously.
The effect of batching is correspondingly large: If an exchange sends five standard transactions with one input and two outputs each, it must create five signatures which require a total of 1100 bytes. If it sends only one transaction with one input and six outputs, it has to sign only once, which is why it uses only 344 bytes. That is not even a third.
The costs for transactions thus decrease when using a platform that bundles transactions. In principle, all platforms that manage coins for their users can offer such a procedure. Despite the reasonably widespread use of Batching, there is still room for improvement – which is a pretty good prospect for the scaling of Bitcoin.
by Christoph Bergmann
This article was originally published in German on Christoph’s blog Bitcoinblog.de, the leading German blog on Bitcoin and other cryptocurrencies. Translated by Aaron Koenig.