TPS is a bad standard of evaluation

Editor’s note: This article from the etheric Fang enthusiasts (ID:ethfans) author: Nic, Carter, translation & proofreading: IAN LIU & Elisa, the daily planet, unauthorized release.

The original link:

We usually take the daily trading volume or “TPS” (transactions per second), to measure the bitcoin network. Many new blockchain often emphasize their network trading volume can reach trillions of TPS, and to ridicule the bitcoin network only 3~4 TPS (throughput in the day 200 thousand to 350 thousand transactions between floating).

But such a comparison is very one-sided. In the evaluation of the whole value transfer system, we should consider three important variables, and the transaction throughput is just one of them:

  • Transaction throughput (TPS)

  • The general characteristic of transaction (transaction size)

  • Clearing guarantee

The transaction throughput and the average transaction size determines the system’s economic productivity, in the financial system is used to measure the flow per unit time transfer; clearing guarantee tells us what the system belongs to the type: can you confirm the transaction fails, do not encounter fraud? Whether the transaction such as cash transactions as timely settlement? Or like a credit card, a 90 day refund period?

Although bitcoin is virtual, but it is more like a clearing system entities, such as credit card and non delayed settlement system of this kind. Bitcoin is a virtual value of the carrying tool: if you have not spent output (UTXO) of the private key, then you are the owner of the money, and have the power to carry out any operation on it; this is not a real asset liability. From this point of view, bitcoin is more like real cash or gold. If you lost bitcoin or is sent to the wrong address, this method had no money to get back. Of course, in some cases, the system can recover the misuse of the money is a good thing; and the cost of credit exists, it is in order to deal with these transactions or estoppel is fraud. The treatment of credit fraud and recovery cost, eventually to fee form transferred to end users, this is why we cannot usually cause the credit card payment is less than 1 or 2 dollars.

So, how do we use consistent criteria to compare different value transfer system, such as bitcoin, Paypal, Visa, SWIFT, and cash? I hope that you will extend to the attention characteristics in addition to the transaction throughput outside, because this is a system of economic productivity one of the factors.

For different value transfer system, the following figure shows the comparison method of ideal: vertical and horizontal axis represent the number of transactions and the average transaction size (amount). In short, the payment network composed of many small transactions, and clearing network by number small but large amount of transaction composition (in US dollars).

 TPS is a bad evaluation standard

We should pay attention to, is under the coverage of each point: This is the average transaction size (amount) * the number of transactions, product per unit time through the system of value.

To further simplify, if you collect different network data, you will find that they tend to congregate in the upper or right end.

 TPS is a bad evaluation standard

Of course, I was too simplified, but I believe you can understand my meaning. Now we will make some actual figures inserted into the ideal model, including entity cash transactions, credit / debit transactions, mobile payment, P2P mainstream digital currency trading, gold trading settlement, clearing and other infrastructure and financial systems (ACH, Fedwire, CHIPS, SWIFT) of the transaction. In my model, into the actual digital chart looks like a more exaggerated L shape, because these systems change in two axes is very intense.

 TPS is a bad evaluation standard

Data from appendix

In the upper left corner, we see Visa cash and credit / debit transactions, they have a single transaction volume greatly and relatively small; in the lower right corner, we can see a clear network giant, which is derived from the London gold market trading settlement, it has the largest single transaction. Reminder: SWIFT is a news service, so the value is not really through the system, here’s a list of the SWIFT data is mainly theoretical data.

Next, let us look at the log view, in order to put all the contents into the.

 TPS is a bad evaluation standard

– the blue dotted line best fitting index

This is the most comprehensive perspective. It was clearly the Ming and Qing Dynasties settlement network in the lower right corner, payment network in the upper left corner. Bitcoin is more and more close to the lower right corner, because the value of the single and higher transaction basis. Next, we first exclude clearing network, to focus the discussion on payment network.

 TPS is a bad evaluation standard

In order to cut the chart shows a subset of data –

As you can see, the vast majority of encryption currency in the X direction is almost no fluctuation — no trading volume, or that they could not reach the mainstream payment network throughput.

So, what exactly is bitcoin?

As the chart shows, bitcoin single transaction size has increased trend. We are very difficult to get accurate numbers, but the average transaction amount can reach thousands or even Shushuwan dollars, the median of more than $100.

 TPS is a bad evaluation standard

– excerpt from the “new economic system: bitcoin,” this is my 2018 Honeybadger in the Baltic Sea Conference Report

Some large transactions from the actual needs: because many will often exchange cash flow into a single transaction, in order to reduce pay; a digit transaction fee means from the fee perspective, large transactions it cost-effective.

More importantly, bitcoin is provided by the world final clearing services, this is not the low value transactions need. We use the system, security system and should provide. It can be said that you won’t be driving an armored personnel carrier on the street, only to buy a pack of cigarettes, because the security you don’t need so much; you don’t wear bulletproof vests to work; you will not put all the cash in the safe, all open to get money or buy food every time, etc..

If we only pay attention to TPS, then bitcoin critics often ignore the fact that the users of the system is very clear what they need. Venmo, Paypal or Visa to protect the low risk of small transactions can be reversed, and this is a good thing, of course, that no bank account is not used; in fact for people without bank accounts, the existing financial facilities almost has no effect. We need to spend a lot of time to establish such a system, just as the establishment of the user to the bitcoin system like faith.

Like bitcoin this broadband but low transaction value big system, suitable for multijurisdictional (cross-border transactions), lack of trust, and require fast clearing transactions. If the trust already exists, can accept Yan Shiqing settlement, or pay the amount is very small, then the general bitcoin trading is not for you! Vanilla Bitcoin is very useful in the treatment of specific time and small transactions accounted for the settlement, but not discussed here!

Bitcoin is located in size / frequency map of the lower right corner of the part.

Bitcoin transaction system 3~4 per second speed, is one of the reasons why bitcoin can continue to attract users. This unique design is not to follow the example of Visa, Paypal, Alipay, or Venmo. They are not comparable in the objective aspect, because they are not doing the same thing.

Then they are compared to the fair? Still fair, but you must do each measure system.

“Why can’t compare between fruit?” – Lil Dicky

Moreover, bitcoin has a near real-time clearing capacity, provides a huge design space to developers: how the clearing system plus a cheaper and more convenient payment in network delay. This also leads to the lightning network and other Layer-2 solutions: cheap and timely transactions, and periodically clearing in the base layer, because not all transactions require settlement in place after a minute; for small transactions, in class is a better choice. Bitcoin — and then the etheric square, in the development of serious thinking in this direction.

In a word, the value transfer system will change along at least three axially, rather than just a. For some people say that “our system can achieve 500 thousand TPS” response should be: what is your cost? The delay in clearing? Whether to use a single verifier? Your transaction system is part of the dollar in the financial system?

My personal point of view, the economic productivity per unit time through the system of value (US $), is a more comprehensive comparative perspective than TPS. The following measurements of each point contain the area, not just the Y axis value; if from a more comprehensive perspective, bitcoin distance Visa only one order of magnitude difference.


The data collection part, we found some new data source. On the main currency from data encryption. The number of transactions and the adjustment is based on 30 days of smoothing, and to “years” to calculate. Currency trading data show that the initial encryption noise and big error in the map, so Coinmetrics transparent adjustment and noise reduction, which makes the performance of the data belonging to a more conservative way in figure.

Some of the old financial system data can be found in the 10-K report, newspapers and magazines, academic papers and reports. Find the CHIPS and Fedwire data are outdated, therefore should be regarded as a rough estimate. SWIFT data from the latest quarterly report and annual report, but this can not be said clearing quantity directly, and can only be regarded as financial information.

 TPS is a bad evaluation standard

Leave a Reply

Your email address will not be published. Required fields are marked *