Won't Bitcoin block size be resolved through simple market economics?

[ranochigo]

If it takes too long to download a block, you could potentially be accepting a transaction that relies on inputs that have already been spent. In other words, it opens you up to double spend attacks.

Are you talking about the possibility of forks being formed from the slow propagation of blocks? Compact block actually speeds up the process significantly in this regard, to the point which I don't think there would be too much delay from propagation of larger blocks. Given that the block intervals are 10 minutes and that there is no need for excessively large blocks, it wouldn't be too much of an issue.

The UTXO set and number of on-chain users is constrained by the maximum block size. You could also say that the UTXO set is a factor of actual book sizes.

Having a smaller block size entails that transactions would take longer to be included in the block. If a user is willing to wait, then block size becomes a non-factor with the size of UTXO but yeah, I get that there is some correlation.

It is trivial to gain access to IP addresses. An attacker could use a single server that controls many IP addresses.

I think it wouldn't be difficult to configure Bitcoin Core to ban an entire subnet. Purchasing many addresses across different subnets would be cost ineffective.

[pooya87]

If it takes too long to download a block, you could potentially be accepting a transaction that relies on inputs that have already been spent. In other words, it opens you up to double spend attacks.

Are you talking about the possibility of forks being formed from the slow propagation of blocks? Compact block actually speeds up the process significantly in this regard, to the point which I don't think there would be too much delay from propagation of larger blocks. Given that the block intervals are 10 minutes and that there is no need for excessively large blocks, it wouldn't be too much of an issue.

There is also the headers first approach where you download the header of the new block so modern nodes would be aware of any new blocks that are found, a fork or otherwise, and the size is simply 80 bytes.
Regardless of all this, even if it takes longer to download blocks there still is no risk, it just takes longer. For instance instead of finding out about the fork in 5 seconds you find out about it in 30 or 45 seconds.

[PrimeNumber7]

If it takes too long to download a block, you could potentially be accepting a transaction that relies on inputs that have already been spent. In other words, it opens you up to double spend attacks.

Are you talking about the possibility of forks being formed from the slow propagation of blocks? Compact block actually speeds up the process significantly in this regard, to the point which I don't think there would be too much delay from propagation of larger blocks. Given that the block intervals are 10 minutes and that there is no need for excessively large blocks, it wouldn't be too much of an issue.

Not necessarily forks. If it takes a long time to download (and verify) a block, it becomes more difficult to quickly determine if a transaction is valid or not. This would make it difficult to even evaluate the risk of accepting an unconfirmed transaction.

It is trivial to gain access to IP addresses. An attacker could use a single server that controls many IP addresses.

I think it wouldn't be difficult to configure Bitcoin Core to ban an entire subnet. Purchasing many addresses across different subnets would be cost ineffective.

The cost of purchasing two IP addresses from different subnets would be the same as purchasing two IP addresses on the same subnet. This might be slightly more complex from a technical perspective if the same VPS is going to be associated with both IP addresses.

Banning an entire subnet would potentially also ban some legitimate users.

Banning an entire subnet would probably be something most node operators would want to do manually. Until the node operator discovers the high number of IP addresses acting maliciously, the attacker would be able to execute a DOS attack. If the banning of a subnet would be done via automation, an attacker could potentially buy rent IP addresses from many subnets that are sufficient to get nodes to ban the subnet, and cause the nodes to be more vulnerable to sybil attacks.

[Wind_FURY]

Q: "Bitcoin can only handle 7 transactions a second. Surely this makes it unusable for day to day transactions?"

A: "The block size is one of the main components limiting transaction throughput. Miners receive a reward for each block: this reward is composed of a subsidy, originally 50BTC per block, currently 6.25BTC per block; and fees for each transaction paid for by individual users. Over time the reward will become so insignificant that the miners will collectively be incentivised to focus on increasing the amount of fees they receive for each block, encouraging them to reach consensus on increasing the number of transactions in each block. This has already happened with segregated witness; other techniques will no doubt be implemented to further increase this, possibly even just increasing the permitted data size of a block. So I believe this problem will just solve itself over time and isn't something to worry about long term."

Over time, off-chain/layer 2 networks will be perfected and built on top of Bitcoin. In a perfect world, if latency would not be an issue, if externalities of an increase in block size would not be an issue, or if a hard fork to bigger blocks would not be an issue, yes, the market should decide how big, or small the block size is. BUT, if there are technical issues that exist, which there are, then it's best to regulate the block size.

[pooya87]

This would make it difficult to even evaluate the risk of accepting an unconfirmed transaction.

Accepting unconfirmed transactions are always very risky whether with highest internet speed or with a dial up, whether with small or big block size.

[ranochigo]

Not necessarily forks. If it takes a long time to download (and verify) a block, it becomes more difficult to quickly determine if a transaction is valid or not. This would make it difficult to even evaluate the risk of accepting an unconfirmed transaction.

Should any capacity increase happen, I would think that the priority would be to strike a balance between the time needed to download and verify the block while still ensuring the security. As of now, 1MB blocks take less than a second to download and verify for me. Compact block** already takes care of most of the heavy lifting required to validate a block.


** Which I think would address the concerns regarding bandwidth until blocks gets exceedingly big and also assuming that internet speed doesn't get faster which is usually not the case.

Banning an entire subnet would potentially also ban some legitimate users.

Agreed.

Actually, why would DOS be a factor with increasing block size in the first place? If the node only requests for the full block after validating the block header and ensuring that it meet the target, the small block header wouldn't be enough to successfully DOS someone.

[No_2]

What I can glean from the responses is that it's not worth increasing the number of TXs per bock because this will significantly increase rate of growth of the Bitcoin blockchain.

[Wind_FURY]

What I can glean from the responses is that it's not worth increasing the number of TXs per bock because this will significantly increase rate of growth of the Bitcoin blockchain.

A small part inside of me wants Bitcoin Cash and its fork Bitcoin Cash SV to have exponentially increased transaction volumes for a decade just to prove that point.

[odolvlobo]

What I can glean from the responses is that it's not worth increasing the number of TXs per bock because this will significantly increase rate of growth of the Bitcoin blockchain.

You can only conclude that it's not worth it by directly comparing costs and benefits. What is the benefit of increasing the throughput? What is the cost of significantly increasing the rate of growth of the block chain? Produce some concrete analysis and then you can come to a conclusion.

[pooya87]

What I can glean from the responses is that it's not worth significantly increasing the number of TXs per bock because this will significantly increase rate of growth of the Bitcoin blockchain.

That's wrong because you are missing an important keyword which I added with a bold font.
We've already increased the number of transactions per block 4 years ago with introduction of SegWit and with Taproot and if some portion of the MultiSig transactions move to it, we will increase it more.
We still need to increase the on-chain capacity for bitcoin to continue to work, specially if we want second layer to be a viable option.

The problem is with the size of the increase, we want to come up with a reasonable capacity increase to still have both scaling and decentralization at the same time. Keep in mind that is not just about the size of blockchain although that is an important factor, it is also about the cost of validating each block.
Increase the size by a lot and you risk centralization and a useless bitcoin, increase it by a little and you risk lack of capacity, mempool congestion, extremely high fees and an unusable bitcoin.

[Wind_FURY]

What I can glean from the responses is that it's not worth significantly increasing the number of TXs per bock because this will significantly increase rate of growth of the Bitcoin blockchain.

The problem is with the size of the increase, we want to come up with a reasonable capacity increase to still have both scaling and decentralization at the same time. Keep in mind that is not just about the size of blockchain although that is an important factor, it is also about the cost of validating each block.

Between increasing throughout/capacity per block, and maintaining decentralization/the network’s capacity to scale out, I believe it’s better to be conservative, and let the hardware capacity catch up. Satoshi may have invented Bitcoin too early.

[d5000]

OP, there is also another problem with your reasoning:

[...] Over time the reward will become so insignificant that the miners will collectively be incentivised to focus on increasing the amount of fees they receive for each block, encouraging them to reach consensus on increasing the number of transactions in each block.

The incentives are not so simple. The transaction fees the miners can get increase principally when the block space is scarce.

Take December 2017 as an example: There were less transactions per block than now (because Segwit transactions weren't still that popular, and legacy transactions take more "weight" in the blocks), but the competition for block space drove the transaction fees to up to 10 BTC per block, which was almost what a miner could get as a block reward (12,5 BTC).

So the miners are, in fact, incentived to keep at least some scarcity.

To show another example, let's look at LTC. They have a "capacity" of about 4 times of Bitcoin's block capacity, and about half of BTC's transactions per hour, but the block rewards for miners are insignificant (much less than 1/10 of Bitcoin's) because there is no scarcity, a transaction will be included even with the minimum fee. (The same happens with BCash).

LTC and BCash could run into serious problems in the future, when the block rewards go down, if they don't achieve to create a "fee market", and this may mean they will need to restrict block space more, if the transaction number isn't increasing. At least their security could enter a long downward trend.

[TangentC]

OP, there is also another problem with your reasoning:

[...] Over time the reward will become so insignificant that the miners will collectively be incentivised to focus on increasing the amount of fees they receive for each block, encouraging them to reach consensus on increasing the number of transactions in each block.

The incentives are not so simple. The transaction fees the miners can get increase principally when the block space is scarce.

Take December 2017 as an example: There were less transactions per block than now (because Segwit transactions weren't still that popular, and legacy transactions take more "weight" in the blocks), but the competition for block space drove the transactions to up to 10 BTC per block, which was almost what a miner could get as a block reward (12,5 BTC).

So the miners are, in fact, incentived to keep at least some scarcity.

To show another example, let's look at LTC. They have a "capacity" of about 4 times of Bitcoin's block capacity, and about half of BTC's transactions per hour, but the block rewards for miners are insignificant (much less than 1/10 of Bitcoin's) because there is no scarcity, a transaction will be included even with the minimum fee. (The same happens with BCash).

LTC and BCash could run into serious problems in the future, when the block rewards go down, if they don't achieve to create a "fee market", and this may mean they will need to restrict block space more, if the transaction number isn't increasing. At least they security could enter a long downward trend.

LTC miners also merge mine Dogecoin at the same time, and it has even higher capacity for transactions than LTC.
So the LTC/Doge miners are profiting from 2 networks at the same time.
So your concerns for their future is unwarranted.

As far as Bcash, the very same miners mining Bitcoin, are also mining BCH.
So BCH will survive as long as it and bitcoin use the same mining infrastructure.

* Average Transaction fees on April 16.
Bitcoin  $27    &   BCH  1 cent
Doge    $1       &  LTC  5 cents

Only the rich will bother to transacts BTC directly onchain,
everyone else will use the alts or a 3rd party services to transact offchain with bitcoin to avoid the high fees.

[pooya87]

As far as Bcash, the very same miners mining Bitcoin, are also mining BCH.
So BCH will survive as long as it and bitcoin use the same mining infrastructure.

This is not merge mining for the miner to be able to mine both at the same time, they'll have to dedicate valuable computing power and electricity to ONLY one cryptocurrency. And they do it for the chain that gives them more profit and also security (ie. it is not going to disappear all of a sudden due to whatever reason including 51% attack like last times).
Those miners that mine both are actually the centralized owners of BCH network that have full control over it (ie. controlling more than 90% of the total hashrate).

[d5000]

LTC miners also merge mine Dogecoin at the same time, and it has even higher capacity for transactions than LTC.
So the LTC/Doge miners are profiting from 2 networks at the same time.
So your concerns for their future is unwarranted.

I think you didn't understand my "concerns". The concerns are related to the fact that mining LTC (and Doge, and whatever other currency with "big blocks" or "high capacity") could become less attractive for miners when the regular block rewards go down. If block space is not scarce, the transaction fee evolution tends to be a race to the bottom - if miners tried to create a cartel and accept only transactions over a certain minimum, for example, it is always rational for some of them to include those which pay less fees than the minimum in their blocks, because their profit will be higher.

Miners don't want high capacities but high fees, which isn't the same, because with higher capacities, the fees would tend to be lower. Even worse, miners profit a lot from users "panicking" about being included in blocks with very high fees (those that pay >100 sat/byte today), and this situation would occur almost never in a bigblocker blockchain.

The "ideal situation" for miners would be a blockchain which stores a high amount of value but at the same time has a limited capacity for transactions, so the fee market is always working. This is however not the ideal situation for users, they may prefer blockchains with low fees like LTC (but the drawback is their security). It is a pretty complex incentive game, because miners must also take care that users don't transfer too much value from Bitcoin to altcoins. So they could tolerate second layer solutions even if they lower the fee a bit.

Perhaps one could say that miners would like "the maximum of capacity which doesn't sacrifice block space scarcity." This could lead indeed to a push for a higher capacity in the far distant future, but not now, now I think the capacity is near the ideal for miners because there is some competition for fees and often more than 1 BTC of fees per block.

But anyway the miners aren't absolutist rulers, so they can't set the rules if the other sectors not agree. "Economic" nodes (those using BTC to pay) are strong in Bitcoin's power game, too, in addition to groups of developers themselves if their expertise is recognized. Both economic nodes and miners could pay their own developers but a group of developers like the Core team is recognized in the community and can justify their decisions (about the rules), then it would be difficult to "overrule" them, like the Bitcoin Unlimited conflict in 2016/17 showed.

And yes I think pooya87 is right - BCash/BTC is not merged mined, miners must decide to which chain they dedicate their hashpower.

[TangentC]

LTC miners also merge mine Dogecoin at the same time, and it has even higher capacity for transactions than LTC.
So the LTC/Doge miners are profiting from 2 networks at the same time.
So your concerns for their future is unwarranted.

I think you didn't understand my "concerns". The concerns are related to the fact that mining LTC (and Doge, and whatever other currency with "big blocks" or "high capacity") could become less attractive for miners when the regular block rewards go down. Transaction fees without scarcity tends to be a race to the bottom - if miners decided to accept only transactions over a certain minimum, for example, it is always rational for some them to include those which pay less fees than the minimum in their blocks, because their profit will be higher.

Miners don't want high capacities but high fees, which isn't the same, because with higher capacities, the fees would tend to be lower. Even worse, miners profit a lot from users "panicking" about being included in blocks with very high fees (those that pay >100 sat/byte today), and this situation would occur almost never in a bigblocker blockchain.

The "ideal situation" for miners would be a blockchain which stores a high amount of value but at the same time has a limited capacity for transactions, so the fee market is always working. This is however not the ideal situation for users, they may prefer blockchains with low fees like LTC (but the drawback is their security). It is a pretty complex incentive game, because miners must also take care that users don't transfer too much value from Bitcoin to altcoins. So they could tolerate second layer solutions even if they lower the fee a bit.

Perhaps one could say that miners would like "the maximum of capacity which doesn't sacrifice block space scarcity." This could lead indeed to a push for a higher capacity in the far distant future, but not now, now I think the capacity is near the ideal for miners because there is some competition for fees and often more than 1 BTC of fees per block.

But anyway the miners aren't absolutist rulers, so they can't set the rules if the other sectors not agree. "Economic" nodes (those using BTC to pay) are strong in Bitcoin's power game, too, in addition to groups of developers themselves if their expertise is recognized. Both economic nodes and miners could pay their own developers but a group of developers like the Core team is recognized in the community and can justify their decisions (about the rules), then it would be difficult to "overrule" them, like the Bitcoin Unlimited conflict in 2016/17 showed.

And yes I think pooya87 is right - BCash/BTC is not merged mined, miners must decide to which chain they dedicate their hashpower

Miners want high fees , Users want low fees.
High Capacity blockchains allow miners to offer lower fees while earning more.
In the end, the Users will choose based on their personal situations and the miners must capitulate or be left behind.

IMO,
Bitcoin is rare , in the fact, they don't really care for mass usage, their goal is corporate usage, and the normal person will never use bitcoin as many here originally believed, it will be the hoard of gold that the corporate users point to as their proof of value, while offering alternatives such as LN or Liquid or their own custom offchain approach, but make no doubt, the normal user will purchase the offchain note separately and never even own bitcoin onchain, that is the reason the blocksize is being limited.
Whether this was the main goal or just the unintended consequences , none the less, it is what is happening.

Can you afford to mine , bitcoin?  Only if you are rich.
Can you afford to buy 1 bitcoin?  Only if you are rich.
Will you use bitcoin to pay local bills? Only if you are rich and like tacking on $27 gift to the miners on every bill you pay

In a normal rational world, bitcoin would have had to lower transaction fees or risk collapse already,
but as time has proven, logic and reason have left the building, global manipulations is ruling the day and it points to bitcoin price forever increasing but only the rich able to afford to own or use it, all others will be forced to use offchain proxy notes.

Correct BTC and BCH are not merge mined like LTC/Doge,  BTC/BCH miners do have a switching systems based on profit.
But they do switch when profit favors, so that is why BCH won't die as long as they can share infrastructure with BTC.
https://ieeexplore.ieee.org/document/8835333

miners have had a choice between BTC and BCH mining because they have compatible proof-of-work algorithms.
Therefore, they can freely choose which coin to mine for higher profit, where the profitability depends on both the coin price and mining difficulty.
Some miners can immediately switch the coin to mine only when mining difficulty changes because the difficulty changes are more predictable than that for the coin price, and we call this behavior fickle mining

[Wind_FURY]

I believe it’s better to be conservative, and let the hardware capacity catch up. Satoshi may have invented Bitcoin too early.

The hardware and internet already catch up, you can run Bitcoin full node on Raspberry Pi (and other ultra low power/cheap computer).

It has not caught up enough that the “obsolete hardware” you can pick up from a swap meet is more than the recommended specification to run Bitcoin Core, and the cheapest bandwidth can do the IBD in mere hours. Think what if Bitcoin was invented 100 years from now.

[Wind_FURY]

It has not caught up enough that the “obsolete hardware” you can pick up from a swap meet is more than the recommended specification to run Bitcoin Core, and the cheapest bandwidth can do the IBD in mere hours.

Can you give example of "obsolete hardware"? Raspberry Pi is cheap, even compared with used 10 years old PC. IBD problem won't be solved anytime soon as long as many internet provider have strict cap/limit.

“Obsolete” for the people of the future’s standards. Probably people of the 2nd-Generation-Bitcoiners. I would believe those people who would be a 40 year old Bitcoiner of today’s grand children. Simple estimate, 20 to 30 years from now.

[d5000]

Miners want high fees , Users want low fees.
High Capacity blockchains allow miners to offer lower fees while earning more.
In the end, the Users will choose based on their personal situations and the miners must capitulate or be left behind.

Please at least try to understand what I write. This is my last post on the issue (at least if there are no new on-topic arguments), we're derailing the thread otherwise.

"High capacity blockchains" will always have low fees, and thus low tx fee earnings for the miners, because miners can't force anybody to pay high fees. To achieve a high security level, they would have to offer high (and constant) block rewards - which leads to more supply inflation and a lower usability as "store of value".

Bitcoin has already decided that the block reward will be decreased every ~4 years. So the profit for miners will come, in a few years, increasingly from transaction fees (like the OP correctly wrote).

The OP has reasoned, thus, that miners would like to increase their fee earnings. But his conclusion, that they could do this lobbying for increased block capacity, is partly wrong because more transaction capacity does not necessarily mean more fees, and both indicators can be even negatively correlated.

Miners could lobby for a very slow increase of block capacity, in a way scarcity is not lowered (for example, an increase of 1% per year or so), but this would become rational only when the fees really threat to lead into a mass exodus to other coins, in the distant future when situations like in late 2017 (with 10$+ fees even at weekends) become normal. I think this point is still far ahead, we have LN now which can be used for lots of scenarios where you wouldn't want to pay high fees.

I refrain from commenting the part about "manipulation" and "logic and reason".

[TangentC]

Miners want high fees , Users want low fees.
High Capacity blockchains allow miners to offer lower fees while earning more.
In the end, the Users will choose based on their personal situations and the miners must capitulate or be left behind.

Please at least try to understand what I write. This is my last post on the issue (at least if there are no new on-topic arguments), we're derailing the thread otherwise.

"High capacity blockchains" will always have low fees, and thus low tx fee earnings for the miners, because miners can't force anybody to pay high fees. To achieve a high security level, they would have to offer high (and constant) block rewards - which leads to more supply inflation and a lower usability as "store of value".

Bitcoin has already decided that the block reward will be decreased every ~4 years. So the profit for miners will come, in a few years, increasingly from transaction fees (like the OP correctly wrote).

The OP has reasoned, thus, that miners would like to increase their fee earnings. But his conclusion, that they could do this lobbying for increased block capacity, is partly wrong because more transaction capacity does not necessarily mean more fees, and both indicators can be even negatively correlated.

Miners could lobby for a very slow increase of block capacity, in a way scarcity is not lowered (for example, an increase of 1% per year or so), but this would become rational only when the fees really threat to lead into a mass exodus to other coins, in the distant future when situations like in late 2017 (with 10$+ fees even at weekends) become normal. I think this point is still far ahead, we have LN now which can be used for lots of scenarios where you wouldn't want to pay high fees.

I get what you are saying, as far as increased blocksize may not increase fee earnings for the miners,
if the number of transactions does not increase to use the capacity.

But in normal economics , higher prices always decreases usage , while lower prices might not increase usage it never decreases it.

Forcing an artificial scarcity to drive transaction prices higher , is more fraud than function.
LTC 4X BTC capacity while Doge ~10X BTC capacity

IMO,
BTC could easily offer a normal non-segwit 10MB block every 10 minutes with no real problems , right now.
And let's be clear when you increase the blocksize capacity to 10 MB per block, if there are not enough transactions,
then it only creates a 1 MB block, as blocksize is adaptive not static.
So the fears of massive blockchain size increases are misplaced.
It will only grow to match the transactions made.
What happens now is when the transaction fee get too high , people just stop using it until fees lower, so that is placing a cap on BTC utility as a way to transfer value.  LN means billions of transactions offchain , that miners will earn nothing, which is why they keep BCH alive.
Miners did lobby for a higher blocksize during the segwit arguments, they were ignored and threatened by Core Devs with an algorithm change that would brick their ASICs. Which is why, you don't hear a peep from them anymore about increasing BTC blocksize.
* As far back as 2010, a 7mb block size was being discussed and might have been implemented then had satoshi not disappeared. *


FYI:
https://www.quora.com/How-much-data-is-consumed-if-I-watch-YouTube-for-1-hour-at-144p
3 years ago,
Watching a 1080P Youtube video consumed 27.61MB in 1 minute, and bitcoin devs claims 10MB in 10 minutes is beyond it.