Reponse to Roger Ver's "Time to End the Block-Size Blockade" essay

[da2ce7]

Hello, many here know of Roger's essay published on the reputable website of the Foundation for Economic Education:

https://fee.org/articles/time-to-end-the-block-size-blockade/


I wrote a considered response to Roger's essay, however, unfortunately, I have not had any response, either from Roger; or from the community at-large. Hence I'm kindly asking for some commentary / responses from the people here on this forum. In particular, are there any part of my post that are wrong, invalid, or misleading?


Original Reddit Post Here

There are many issues with this essay; really I was quite disappointed with Roger. The primary one being intellectual dishonesty from the use of a bad analogy. While this in itself doesn’t invalidate Rogers other claims, it doses put the other claims in poor light.

Firstly, to address the fundamental issue with the Starbucks analogy: Bitcoin isn’t like Starbucks Corporation; while there is only one Bitcoin Blockchain, there are tens of thousands of Starbucks stores.

To respond to increases in demand Starbucks can just open more stores (that operate more-or-less independently). However, with Bitcoin, we cannot just ‘add more nodes’ to respond to an increase in demand; as every (full) node must carry the entire burden of the network.

Roger dose correctly state:

Any time a Bitcoin user is willing to pay a fee that is larger than the marginal cost of including the transaction in a block, it makes economic sense for a miner to include it.

However, his statement is intellectually dishonest as he doesn’t analyze what the marginal cost actually is. Where in reality, the marginal cost of including a transaction in a not-full block is virtually zero. Meaning that it doesn’t matter what size blocks you have, they will constantly be full. (with exception to miners who have differing policies, but the economically rational miner will include transactions that have any fee above zero).

As we continue Roger doesn’t analyse the problems and benefits of blocks being full, instead, based upon his Starbucks fallacy, he just assumes that full blocks are bad. There are two main arguments of why ‘full blocks’ isn’t the evil that Roger plays it to be:

• It is economically unavoidable. (The marginal cost of creating and including transactions is virtually zero).
• Without full blocks there isn’t any fee-pressure. Where fees are increasingly important to the long-term health of the network. As the subsidy runs down, this should be compensated by an increase in transaction fees.

Yet neither of these arguments are addressed in Rogers essay.

Thus we come to the core contention of Rogers essay: (I paraphrase) “If we increase the blocksize, blocks will meet capacity again”. This is namely false, as it is completely reasonable to expect that blocks of any size to be full.

Roger dose a good job of arguing how is it is ‘possible’ to increase the blocksize. He completely fails to put it into a well-reasoned context of ‘why’ we need to rise the blocksize limit.

The marginal cost of getting into the ‘full’ blocks that are on the network now (with the 1MB limit), is only 8 cents (US). see: https://bitcoinfees.21.co/

• Roger dose not address that there is virtually no fee-curve meaning: There are only really two types of transactions on the network: spam and fee paying. (edit: in that to outbid the spam, you only need to increase your fee from 4 cents to 8 cents; suggesting there isn't a strong demand for extra block-space from real fee-paying users.)
• Roger also doesn’t address the economic implications of these 8 cent fees on the network. He doesn’t address at what point do fees become so large that it does have significant economic implications.
• Roger doesn’t address any argument of what point we should start considering a fee market important to the future of Bitcoin (with miners being paid by fees).

Overall this essay is an embarrassment, and anti-intellectual. I can only assume that Roger didn’t put this essay out to rigours peer-review before publishing it.

[smoothie]

Not surprising. This is the same guy who publicly came out 6-8 months before MTGOX went belly up and said there were no financial problems with MTGOX because he "read letters from banks and attorneys".

[AliceGored]

Hello, many here know of Roger's essay published on the reputable website of the Foundation for Economic Education:

https://fee.org/articles/time-to-end-the-block-size-blockade/


I wrote a considered response to Roger's essay, however, unfortunately, I have not had any response, either from Roger; or from the community at-large. Hence I'm kindly asking for some commentary / responses from the people here on this forum. In particular, are there any part of my post that are wrong, invalid, or misleading?


Original Reddit Post Here

There are many issues with this essay; really I was quite disappointed with Roger. The primary one being intellectual dishonesty from the use of a bad analogy. While this in itself doesn’t invalidate Rogers other claims, it doses put the other claims in poor light.

Firstly, to address the fundamental issue with the Starbucks analogy: Bitcoin isn’t like Starbucks Corporation; while there is only one Bitcoin Blockchain, there are tens of thousands of Starbucks stores.

To respond to increases in demand Starbucks can just open more stores (that operate more-or-less independently). However, with Bitcoin, we cannot just ‘add more nodes’ to respond to an increase in demand; as every (full) node must carry the entire burden of the network.

Imperfect analogy is imperfect, obviously. But there is something to be learned from the idea that if capacity is fully saturated... any additional demand prices out some of the old users and some of the old use cases, as long as they are willing to outbid their competitors (other Bitcoin users). Yes, increasing throughput increases resource demand on full nodes, but so does segwit, moreso than a simple 2MB upgrade. (3MB increase in potential max load with 0.8MB throughput gain). Yes nodes can ignore segwit and consider it an altcoin, they can prune, they can reject inbound... I'm talking about full archival nodes. It's important to recognize that hardware and internet bandwidth has improved over the last 7 years.

Roger dose correctly state:

Any time a Bitcoin user is willing to pay a fee that is larger than the marginal cost of including the transaction in a block, it makes economic sense for a miner to include it.

However, his statement is intellectually dishonest as he doesn’t analyze what the marginal cost actually is. Where in reality, the marginal cost of including a transaction in a not-full block is virtually zero. Meaning that it doesn’t matter what size blocks you have, they will constantly be full. (with exception to miners who have differing policies, but the economically rational miner will include transactions that have any fee above zero).

Now we get to the heart of the matter. You say that miners would include any non-zero fee transaction just because they can. This is where the philosophical difference comes to a head. Some feel that without a production quota on blocksize, set by a priesthood of Core devs, miners would simply bloat blocks to infinity. They would willingly destroy the independent node network, simply because they're dumb? or they only care about this week's profits?

Others find this argument ridiculous, and counter to the very "core" principles that govern Bitcoin: Where free market incentives, and CPU consensus, rule the system... not centrally ordained software ministers.

Miners are uniquely placed [and designated by satoshi via the consensus mechanism] to best determine the size of their own blocks because they are directly beholden to the market, unlike devs with ambiguously disclosed stock options in VC startups.

As we continue Roger doesn’t analyse the problems and benefits of blocks being full, instead, based upon his Starbucks fallacy, he just assumes that full blocks are bad. There are two main arguments of why ‘full blocks’ isn’t the evil that Roger plays it to be:

• It is economically unavoidable. (The marginal cost of creating and including transactions is virtually zero).
• Without full blocks there isn’t any fee-pressure. Where fees are increasingly important to the long-term health of the network. As the subsidy runs down, this should be compensated by an increase in transaction fees.

Yet neither of these arguments are addressed in Rogers essay.

Thus we come to the core contention of Rogers essay: (I paraphrase) “If we increase the blocksize, blocks will meet capacity again”. This is namely false, as it is completely reasonable to expect that blocks of any size to be full.

I suppose this is the point where you should produce a chart where blocks were constantly at 1MB for the last 6 years... because free globally replicated storage. The reality would probably be more like what we've experienced so far... a steady rise in the use of Bitcoin, and a commensurate increase in blocksize. Not an immediate jump to the new max.

Again with the argument that miners bloat to infinity without a central authority stopping them...  all records of blocksize statistics refute this position.

Roger dose a good job of arguing how is it is ‘possible’ to increase the blocksize. He completely fails to put it into a well-reasoned context of ‘why’ we need to rise the blocksize limit.

The marginal cost of getting into the ‘full’ blocks that are on the network now (with the 1MB limit), is only 8 cents (US). see: https://bitcoinfees.21.co/

• Roger dose not address that there is virtually no fee-curve meaning: There are only really two types of transactions on the network: spam and fee paying. (edit: in that to outbid the spam, you only need to increase your fee from 4 cents to 8 cents; suggesting there isn't a strong demand for extra block-space from real fee-paying users.)
• Roger also doesn’t address the economic implications of these 8 cent fees on the network. He doesn’t address at what point do fees become so large that it does have significant economic implications.
• Roger doesn’t address any argument of what point we should start considering a fee market important to the future of Bitcoin (with miners being paid by fees).

Overall this essay is an embarrassment, and anti-intellectual. I can only assume that Roger didn’t put this essay out to rigours peer-review before publishing it.

• Miners do face a supply curve, and a propagation cost to larger blocks. The distinction of fee paying/spam is non-functional, as zero fee transactions are already ignored by miners and most nodes. Median fees are 8 cents or more, up from 2 cents a couple months ago, a 400% increase.
• He doesn't have to. It's obvious to an honest observer that 8 cents for the smallest tx possible is already greatly diminishing potential use cases. Blockstream says this is fine. Because (not currently available) payment contracts in LN will handle all that.
• It's not really up to Roger, or the most influential developers, to determine the dynamics of a fee market. It would and will exist with miners facing the free market and subsequently setting their own production levels. Beware the dangers of a centrally planned economy.

I could mirror your petulant conclusion, but, meh.

[Lauda]

Overall this essay is an embarrassment, and anti-intellectual. I can only assume that Roger didn’t put this essay out to rigours peer-review before publishing it.

Was something better expected from Ver, the person who has been spreading bullshit ever since the time of Mt.Gox? These articles are just an attempt to gather more support for their HF. It's unfortunate that they go social media instead of actually trying to improve the existing infrastructure.

Yes nodes can ignore segwit and consider it an altcoin, they can prune, they can reject inbound...

No, they can't consider Segwit "an altcoin".

It's important to recognize that hardware and internet bandwidth has improved over the last 7 years.

People keep saying that, yet the number of nodes is on a steady decline. Do I have to note that the resource usage kept growing tremendously over the 7 years?

This is where the philosophical difference comes to a head. Some feel that without a production quota on blocksize, set by a priesthood of Core devs, miners would simply bloat blocks to infinity.

They certainly would not be the first 'group' to spam the network in an attempt to support their vision.

It would and will exist with miners facing the free market and subsequently setting their own production levels. Beware the dangers of a centrally planned economy.

So in a scenario in which miners get paid to set them to abrupt levels, all is fine right?

[AliceGored]

Yes nodes can ignore segwit and consider it an altcoin, they can prune, they can reject inbound...

No, they can't consider Segwit "an altcoin".

Sure they can, in fact, any nodes that don't upgrade to a segwit friendly version will consider pure segwit txs to be an altcoin.

It's important to recognize that hardware and internet bandwidth has improved over the last 7 years.

People keep saying that, yet the number of nodes is on a steady decline. Do I have to note that the resource usage kept growing tremendously over the 7 years?

Check your premises, running the qt client used to be the most common way to interact with the network, now it's not. SPV type interaction is well developed and widespread at this point. 

This is where the philosophical difference comes to a head. Some feel that without a production quota on blocksize, set by a priesthood of Core devs, miners would simply bloat blocks to infinity.

They certainly would not be the first 'group' to spam the network in an attempt to support their vision.

Incoherent non sequitur.

It would and will exist with miners facing the free market and subsequently setting their own production levels. Beware the dangers of a centrally planned economy.

So in a scenario in which miners get paid to set them to abrupt levels, all is fine right?

You strongly favor a centrally planned economy, which is a position, just not one I agree with.

[Lauda]

Sure they can, in fact, any nodes that don't upgrade to a segwit friendly version will consider pure segwit txs to be an altcoin.

That would not make much sense. Please provide a source using the terminology "pure segwit txs" equal "an altcoin".

Check your premises, running the qt client used to be the most common way to interact with the network, now it's not. SPV type interaction is well developed and widespread at this point.  

Exactly who was talking about SPV? Nobody. So you're not denying that the resource usage has increased and thus the number of nodes keep going down.

Incoherent non sequitur.

Sorry, failed attempts at spamming can be easily identified.

You strongly favor a centrally planned economy, which is a position, just not one I agree with.

Not necessarily, I favor the best proposals, and BU is certainly anything but the best proposal.

[AliceGored]

Sure they can, in fact, any nodes that don't upgrade to a segwit friendly version will consider pure segwit txs to be an altcoin.

That would not make much sense. Please provide a source using the terminology "pure segwit txs" equal "an altcoin".

It isn't about terminology, it's about functionality. My non-segwit node doesn't understand anything about segwit. You may use it, but only miners and upgraded nodes understand anything about it. And if you want to pay a non-segwit address, you will not be using segwit. For further research see: Mircea Popescu (admittedly a loon), and his views on your altcoin.

Check your premises, running the qt client used to be the most common way to interact with the network, now it's not. SPV type interaction is well developed and widespread at this point.  

Exactly who was talking about SPV? Nobody. So you're not denying that the resource usage has increased and thus the number of nodes keep going down.

I'm saying correlation =/= causation. SPV has served as an increasingly popular substitute for running a full node, which used to be the main way to use Bitcoin.

[da2ce7]

A couple notes of correction and encouragement to start, you misspelled the title of your post, response=yes, responce=no. Also, try "does" instead of "dose". I understand that english is not everyone's first language, so let's look at the arguments and leave the grammatical competence for now. 

Thank-you for your well-written response; I've updated the title in correction.

Now we get to the heart of the matter. You say that miners would include any non-zero fee transaction just because they can. This is where the philosophical difference comes to a head. Some feel that without a production quota on blocksize, set by a priesthood of Core devs, miners would simply bloat blocks to infinity. They would willingly destroy the independent node network, simply because they're dumb? or they only care about this week's profits?

Others find this argument ridiculous, and counter to the very "core" principles that govern Bitcoin: Where free market incentives, and CPU consensus, rule the system... not centrally ordained software ministers.

Miners are uniquely placed [and designated by satoshi via the consensus mechanism] to best determine the size of their own blocks because they are directly beholden to the market, unlike devs with ambiguously disclosed stock options in VC startups.

This is a very interesting view on the matter; one that wasn't at-all represented in Roger's essay.

When the mining is centralized, or quite a small group running it; it is easy for the miners to voluntary set their own limit via cartel. Hopefully, mining doesn't remain so centralized and it will gradually become reasonable for new entrants to join the mining market once the mining hardware efficiency has hit hard technical limits.

Requiring the use of a cartel to set block size limits, (that are applicable to the mining cartel), locks in the mining and node structure we have now.  As there is a strong competitive advantage for centralizing and having large blocks.

As a user of Bitcoin, I quite like that my node enforces the blocksize limit; I would be very upset for this consensus power to be taken away from me and given to the miners (cartel).

Of course, this is a huge detour from the status quo, the burden of evidence is on the proponents who wish to make such a change.

I suppose this is the point where you should produce a chart where blocks were constantly at 1MB for the last 6 years... because free globally replicated storage. The reality would probably be more like what we've experienced so far... a steady rise in the use of Bitcoin, and a commensurate increase in blocksize. Not an immediate jump to the new max.

Again with the argument that miners bloat to infinity without a central authority stopping them...  all records of blocksize statistics refute this position.

Thank you for this piece of correction.  I really should have said “minimum relay fee-paying transactions”; not “non-zero fee transactions”.

You can see in the average blocksize charts (for example, when looking when Bitcoin had the 'stress test') that virtually every minimum relay fee-paying transaction has been included by the miners (until recently when blocks finally became ‘full’).

The core of my argument still stands, the seen behaviour that miners have been including all transactions that have complied with the minimum-fee-policy that Bitcoin Core has set. I don’t see any reason why this should stop at 2mb or 100mb. The miners generally just use the default policy that Bitcoin Core sets; One would expect the miners to get greedier for fees when fees become a larger % of their revenue.

• Miners do face a supply curve, and a propagation cost to larger blocks. The distinction of fee paying/spam is non-functional, as zero fee transactions are already ignored by miners and most nodes. Median fees are 8 cents or more, up from 2 cents a couple months ago, a 400% increase.
• He doesn't have to. It's obvious to an honest observer that 8 cents for the smallest tx possible is already greatly diminishing potential use cases. Blockstream says this is fine. Because (not currently available) payment contracts in LN will handle all that.
• It's not really up to Roger, or the most influential developers, to determine the dynamics of a fee market. It would and will exist with miners facing the free market and subsequently setting their own production levels. Beware the dangers of a centrally planned economy.

I could mirror your petulant conclusion, but, meh.

1. What!? What does that mean?  The fee curve in bitcoin is flat.  That means that there is very little price discovery for fees; and that most people don't even care about the required fees. (seen by people often overpaying a huge amount for no gain).

The reason that the fee-curve is flat; is that the minimum relay and inclusion policies do not allow it to be developed at the lower fee ranges.

If there was very strong demand, and therefore competition, for the block space, what you would see is many levels of fees.  Say. 8c, 10c, 15c, 30c.  That would all have different expected functionally (average confirmation time).

Yet, now there is really just two levels: (minimum relay), and 8c, immediately confirm.

This, economically, suggests that blocks are not saturated (for real-fee paying users), and that there is very little economic pressure for more transactions to be included.

2. What!? Again. This is a half-truth.  Yes, there are economic use-cases that are placed outside the market from 8c fees. But that says nothing for the economic ‘importance’ of such transactions. Well, it at least says that the said economic use-cases are not economically significant enough to competitively pay 8c transaction fees.

Using the blockchain for storage is a potential use case if the fees were insanely low; but thankfully they are not.

3. It is up to the person promoting the change to explain why this change doesn’t upset the status quo.  It is completely fair to expect a understanding of how blocks are going to be paid for by fees; if the block-size is very large.

It is also reasonable to expect that the blocksize (2MB) that Roger is promoting is the thin edge of the wedge; as he has promoted bigger blocks before on many occasions. The question of the long-term viability of Bitcoin under huge blocks is implicitly at hand.

[Kakmakr]

I have not seen any attempt by Roger Ver to pull the parties together in this community. Instead of doing that, he jumps on the freedom of speech bandwagon and flames this forum with his supporters and then creates a supposedly uncensored forum in direct competition with Bitcointalk.org

It seems to me that he is out to divide the Bitcoin community and this will do more harm than good in the long run. We are getting tired of all this fighting over block sizes and just want Bitcoin to grow. Do something positive for Bitcoin, and take less out of it for yourself. ^hmmmmm^

[RealityTruth]

I think we should roll back the blockchain so that we can have smaller block size again and stolen Mt Gox coins back.

[da2ce7]

@AliceGored

I must admit that going through your history it is clear that you are far, far, more technically and intellectually competent than either Roger, or your average large-blocker.

It would be wise for others to take care when responding to you.

Not that I necessarily agree with what you say.

[AliceGored]

@AliceGored

I must admit that going through your history it is clear that you are far, far, more technically and intellectually competent than either Roger, or your average large-blocker.

It would be wise for others to take care when responding to you.

Not that I necessarily agree with what you say.

I'm flattered. [hurriedly deletes snarky opener above]

Thanks for the engagement on the issues. Things get heated because both sides of this issue care passionately about it... and when you think about it... you can't have light without heat.

[hdbuck]

SPV has served as an increasingly popular substitute for running a full node, which used to be the main way to use Bitcoin.

SPV is cancer.

Running a full node is the only way.

[franky1]

the only way to truly get the full node count up is to make independantly holding and transacting more useful ONCHAIN.

but here is ways to reduce the node count
1. increasing fee's due to bottlenecking the network making people use offchain solutions to get cheaper transactions
2. pruning data automatically makes a node no longer full node
3. offering sidechains. simple making people use alternative chains makes people not want to protect bitcoin and protect the sidechain instead
4. offering LN. after all why protect the blockchain if only 1% of your transactions touch the blockchain.
5. no witness mode. why run a full node if your not going to validate a transaction/block.

now all of those features are what core/blockstream are pushing for. .. which is the funny part. every doomsday they cry about applies to their own features


and there is absolutely no reason at all in the real world why 2mb is going to kill bitcoin.
the only thing anti-HFers can do is exaggerate:
instead of thinking about 2mb, they shout out about 8gb blocks.
instead of thinking about 2mb running on normal computers they shout about data centres..

the truth is that even on a raspberry Pi 3.. and bottom line ADSL internet. bitcoin works fine even with 2mb..
if you think that is completely wrong and 2mb is still bad... then explain segwits 3mb example block data.

oh wait.. you will fail with a rebuttal by mentioning the only buzzword you have left "quadratics".. well guess what.. libsecp256k1 has improved validation times.. not just 2x, not just 3x, not just 4x but 5x....

the only time that quadratics becomes an issue is in regards to LN hub settlement transactions(lots of multisig signatures in one tx).. and guess who wants LN offchain solution...... oh yea, its core/blockstream..
which makes it not really an issue for ON CHAIN REAL CAPACITY GROWTH.

wait lets have another laugh..
"hard forks are bad" so if miners want to hard fork.. core will... wait for it.. they will make their own hardfork and render bitcoin difficulty weak, render ASICS useless (research sha3)

here.. this is adam back self promoting his crap and promoting how he wants to screw bitcoin up
https://en.bitcoin.it/wiki/Hashcash

[gentlemand]

I'm sure our Rog is a lovely fella but I'm not massively inclined to listen to him on such matters.

[Quickseller]

The cost of adding an additional transaction into a block is the incremental chance that a block will get orphaned after a transaction is included in said block. As technology improves, the cost of including the n^th additional transaction will decline, so miners will be able to safely include additional transactions in their found blocks over time.

Miners will not include an additional transaction unless the lowest transaction fee is sufficient to cover the additional risk of a block getting orphaned.

[franky1]

The cost of adding an additional transaction into a block is the incremental chance that a block will get orphaned after a transaction is included in said block. As technology improves, the cost of including the n^th additional transaction will decline, so miners will be able to safely include additional transactions in their found blocks over time.

Miners will not include an additional transaction unless the lowest transaction fee is sufficient to cover the additional risk of a block getting orphaned.

this guy gets it

average internet speed has increased
average hardware has increased (raspi 3 is many times better than raspi model A)
libsecp256k1 is 5x better..

[danda]

The simplest response to Roger's essay is this:

Consensus layer changes only happen if there is widespread consensus that the change needs to happen. Increasing the blocksize is controversial.  It does not have consensus and likely never will.  Therefore, the blocksize limit will not be increased.

[franky1]

The simplest response to Roger's essay is this:

Consensus layer changes only happen if there is widespread consensus that the change needs to happen. Increasing the blocksize is controversial.  It does not have consensus and likely never will.  Therefore, the blocksize limit will not be increased.

by one party not even releasing code to even allow their own devoted users the choice, is the cause of the controversy..

if every implementation had code for a hard fork.. then and only then is it a fair and open opportunity to let the community decide for themselves.
but instead by withholding code, withholding the opportunity for choice and ofcourse causing controversy then ofcourse some people who may want the change, wont bother showing their support publicly, because they understand it cant happen due to one particular party vetoing it out before it has even had a chance for open community choice.

if core/blockstream are so sure that it never will happen. they might aswell release the code and not worry about it because in their eyes no one will use the bigger buffer limit..
the only reason they fear it, is because they know that the community want it. and so are doing all they can to avoid it

lets see what happens when one of their own releases HF code.. will they give in and support it or throw luke JR under the bus just like they did with hearne and gavin..

im thinking they would prefer to do another REKT campaign, but this time on luke JR. but we shall see, soon

[thejaytiesto]

Bitcoin is dead as we know it if you raise the block size. The node count as we have it now is low already at around 5000. If you double this, you can expect a lot of the widespread % to decrease. Even if the total node count went up, the system would be more centralized if specialized services for running nodes arise. Bitcoin is just fucking dead as a censorship resistant tool if guys like Roger Ver were to call the shoots. Lucky it seems we are still safe with Core team., but im tired of the idiots that still shill Bitcoin Classic and the rest of crap.