[cowtools]

>What renders this whole calculation wrong is that it doesn’t account for transaction fees AND transaction fees growth which was also supposed to be growing exponentially, had Bitcoin not been artificially limited to ~ 2 MB blocks every ten minutes

I agree that bitcoin should use an adjustable block size, but I'm not exactly sure that this logic follows. Can you explain this to me? If the network could handle more transactions per block, wouldn't the fees per transaction just decrease?

[intabli]

Correct, the fees would decrease, but when when you can only process 200k transactions every ten minutes, there’s an upper limit of how much people are willing to pay before it becomes absurdly expensive. This reached almost $100 in the 2017 rush. It’s unsustainable and renders the network unusable for most of the planet.

When you on the other hand can process millions of transactions every block (and scalable as needed), you’ll charge far less per transaction while still allowing miners to generate a huge (and ever growing) revenue, while also enabling most of the planet to transact next to free.

From an environmental perspective, if you divide the hash rate environmental impact on a mere 200k transactions every ten minute, the carbon foot print per transaction would seem absurdly high. Do the same calculating for millions of transactions every ten minutes, and you’ll arrive at a far better environmentally friendly figure.

[FabHK]

So currently every transaction costs around $101, of which $1 visible in transaction fees, and $100 invisible in mining reward (aka seignorage, leading to money supply growth of 1.7% p.a.).

To maintain current mining hash power, as the mining reward halves away to zero, every transaction would have to incur around $100 in fees.

> still allowing miners to generate a huge (and ever growing) revenue, while also enabling most of the planet to transact next to free

If blocks would grow 10x, the blockchain would grow about 50GB per month, every human being could do around 10 transactions in their lifetime, and those would cost around $10 each. I don't see how huge and growing revenue for miners is compatible with "next to free" transactions.

[intabli]

I think you will enjoy this great presentation by Peter Rizun and Andrew Stone:

https://youtu.be/5SJm2ep3X_M

[FabHK]

That was an interesting talk, thanks. Looks like with some optimisations they could support blocks of around 100 MB, yielding around 500 transactions per second. Would mean though that the blockchain grows by a few hundred GB every month.

[joshcryer]

Just to clarify your number, an average Bitcoin block (currently) has about 2000 transactions, each of which, be it a transaction for half a cent, or a billion dollars, cost around $3 each. What you described is what happened when Bitcoin was actually being adopted by people, the mempool, that is, transactions not added to the latest block, started to fill up, and the market took hold and people who offered more to transact were included in the block. You don't have to offer more for the transaction and your transaction will sit in the mempool forever. Offering more is simply a way to expedite your transaction.

[_alex_]

Then why isnt bcash or bsv more valuable than bitcoin?

[cowtools]

In my experience, BTC has more "brand recognition" amongst uninformed investors who are more concerned with "number go up" technology than digital cash technology

[intabli]

If you measure value by the thing Bitcoin came to replace (fiat), your vision will be distorted.

If you measure value as a consequence of utility, they’re certainly far more valuable than BTC (BSV however is a corp coin and has weird copyrights so I wouldn’t touch it with a stick).

[hajhatten]

> This reached almost $100 in the 2017 rush.

Ethereum, with a bigger blocksize had transaction fees going up to $3,500 a couple of weeks ago. A larger blocksize is not the answer.

https://web3isgoinggreat.com/?id=popular-nft-mint-spikes-eth...

[intabli]

Ethereum is account based, not UTXO based, has a completely different overhead and you can’t possibly compare it to Bitcoin.

[wmf]

Why wouldn't fees go all the way to zero? Zero times a zillion transactions is still zero.

[mustyoshi]

Miners still have to choose to include your tx. The risk of losing during two blocks being mined goes up as blocksize does (longer to transmit across the network), so you still have to incentivize the miner to include your tx.

Ofc there will be miners who make much larger blocks no matter what, but it comes down to how urgent is your tx?

[shawabawa3]

...because miners choose whether to include a transaction or not. Obviously they have no incentive to include a zero fee transaction

There would be an equilibrium where the fee is worth the marginal cost of mining it

[colinmhayes]

Unless miners start colluding their strategy will become "include every possible transaction with fee greater than $0." which means fees will settle at the smallest possible value greater than 0.

[imtringued]

Yeah sure, everything I buy costs the smallest possible value greater than zero.

shawabawa3 already explained it "There would be an equilibrium where the fee is worth the marginal cost of mining it"

The miner has cost X and wants profit Y. He adds those together and now the price is X+Y. Someone wants his transaction on the blockchain, he is going to pay X+Y unless there is a miner that offers an even lower rate.

You're making the assumption that the block reward is high enough to make transaction fees irrelevant, which is a trivial observation and not what we are interested in talking about here.

[colinmhayes]

There is no cost to adding a transaction to your block if its size is uncapped. Once a miner completes a block why wouldn't they maximize the profit they can make on it?

[imtringued]

If we assume no blockreward then the fees will be whatever motivates miners to include your transaction. If the fee was zero, then no miner would even bother.