The Barry Silbert segwit2x agreement with >80% miner support.

[d5000]

Will segwit reduce the fees someone is willing to pay for a given transaction? Maybe.
Will segwit reduce the average fee when people manually set "high" fees to be at the "top of the list"? No.

And why should the users not reduce fees when block capacity gets higher and blocks are not longer full? The actual user behaviour to (almost panicky) set very high fees is a consequence of the full blocks. You can see that at weekends, average fees are lower then because demand is lower.

You can't have a soft fork and 2 MB.

According to a (veteran) user of the German local forum there seems to be a way, although obviously only "new-style" (P2WPKH/P2WPSH) transactions would benefit from the 2MB "base" size.

[ComputerGenie]

Can you support the claim that fees are proportionally related to size with an equation?

It's simply supply and demand. Supply of capacity lower than the demand of capacity = higher fees. The concept of fee market is retarded, a concept based on devs not understanding economics and/or accountancy. Segwit capacity to accept more transactions can only happen is users use it. If users stick to the current means of transaction, the capacity will be the same.

So, in other words, "no" you can't support your claim, you're just assuming what will happen (and ignoring the possibility that it will have exactly zero impact on fees).

Already have.
No such a thing. Nothing is ever static in economic.

No, you haven't supported your claim, you have asserted assumptions. You are assuming that users will do x if size is y, with no factually mathematical correlation between x and y.
You're conflating "static" and "impact" (as well as conflating "proof" and "belief"). I didn't say that fees would never change, I said that you cannot mathematical show that fees will, in fact, drop as a direct result of segwit.

[ComputerGenie]

Will segwit reduce the fees someone is willing to pay for a given transaction? Maybe.
Will segwit reduce the average fee when people manually set "high" fees to be at the "top of the list"? No.

And why should the users not reduce fees when block capacity gets higher and blocks are not longer full? ....

For the same reason that fees rose in the 1st place, the mentally that higher fees = faster confirmation. Even if you make usable size 25M, there's still more tx than space and people will "fight" for that space through higher fees offered (thus increasing the higher average). Because, math.

[jbreher]

UASF wins, because all the jihancoins will go to 0 due reorg

You User Accidented Segwit Failers are just too precious. You do realize that it is trivial to make such reorgs impossible, right?

[The One]

Can you support the claim that fees are proportionally related to size with an equation?

It's simply supply and demand. Supply of capacity lower than the demand of capacity = higher fees. The concept of fee market is retarded, a concept based on devs not understanding economics and/or accountancy. Segwit capacity to accept more transactions can only happen is users use it. If users stick to the current means of transaction, the capacity will be the same.

So, in other words, "no" you can't support your claim, you're just assuming what will happen (and ignoring the possibility that it will have exactly zero impact on fees).

Already have.
No such a thing. Nothing is ever static in economic.

No, you haven't supported your claim, you have asserted assumptions. You are assuming that users will do x if size is y, with no factually mathematical correlation between x and y.
You're conflating "static" and "impact" (as well as conflating "proof" and "belief"). I didn't say that fees would never change, I said that you cannot mathematical show that fees will, in fact, drop as a direct result of segwit.

Capacity will only go up if all users use segwit. If they don't then capacity will remain the same and the fees will not go down.

Can you support the claim that fees are proportionally related to size with an equation?

I never said the fees will drop.

[dinofelis]

...Well, there remains something unclear about this, and when I ask, I get alternating answers ...

I don't know which one corresponds to what is BIP148.

When in doubt, read the BIP.

While this BIP is active, all blocks must set the nVersion header top 3 bits to 001 together with bit field (1<<1) (according to the existing segwit deployment). Blocks that do not signal as required will be rejected.

Yes, but what does that mean, the block gets rejected ?  Does that mean that the node will wait for another block, that does satisfy the criteria, *and is a direct successor* of the last valid block ?  Or does it mean that these block headers are still accepted, but not "counted" or something ?  Because someone told me that even if there continued to be a single chain with signalling and non-signalling blocks, it would only pick the signalling ones from the chain.

However, if a non-signalling block is truly rejected in full, then the node will wait for a DIRECT SUCCESSOR to the last valid block, and hence, will stop unless there is a true fork in the bitcoin chain, with two prongs: the "legacy" one, and a "pure signalling" one.  Someone told me that that was NOT the case.

[European Central Bank]

You User Accidented Segwit Failers are just too precious. You do realize that it is trivial to make such reorgs impossible, right?

yep, but it also makes an official and irreversible chain split. there ain't no going back.

[dinofelis]

You User Accidented Segwit Failers are just too precious. You do realize that it is trivial to make such reorgs impossible, right?

yep, but it also makes an official and irreversible chain split. there ain't no going back.

This chain split is acted when the first few bip148 blocks are making a prong from the legacy chain, in fact.  And these bit148 blocks with minority hash rate need to get listed as a separate coin on exchanges BEFORE users get in any way their (economic) say, by selling one type of coin, and buying the other type of coin. 

Now, it would seem logical that if two coins are listed on exchanges, that both chains become bilaterally protected so that everything that happens on one chain is not going to be gone in a puf of logic whenever the legacy chain would be overtaken by the forking chain, so as to keep the holdings that people have.  After all, for those *selling* legacy coins, to buy bip148 coins, they couldn't care much ; but for those buying legacy coins, they may want these transactions to be made immutable and would not like to be reorganized away.  So most probably, if the two coins are listed on exchanges, the legacy chain will propose a hard fork to protect itself from re-organisation (the simplest hard fork would be a check point after the split).

[ComputerGenie]

Capacity will only go up if all users use segwit. If they don't then capacity will remain the same and the fees will not go down.

I never said the fees will drop.

Pretty sure that "If [not x] then ...will not go down", implies that fees will go down if x. Because, English.

[ComputerGenie]

...

I could be wrong, "rejected" might mean something other than "not accepted". 

[d5000]

Even if you make usable size 25M, there's still more tx than space and people will "fight" for that space through higher fees offered (thus increasing the higher average). Because, math.

Ah OK, if you part from the assumption that 2+x MB won't be enough for the current transaction volume, then I can agree.

But I don't think so. The present enormous mempool sizes are created because of the backlogs of previous hours and days.

This chart (transactions added to the mempool per seconds) confirms that Segwit for now will be enough even without the 2MB expansion. Average transaction rate (mempool, not blocks!) is about 3,5 per second now - if we think of an "effective" capacity increase to 2,2 MB by Segwit alone that will mean that the blocks - giving today's transaction rate - will be only full to 55-60%, given the current average transaction size of about 500 bytes.

On blockchain.info you can also find charts that indicate that the transaction rate grows less than 50% per year. That would mean that with Segwit alone blocks will be full at the end of 2018, and with Segwit2x you have time until 2020/21. But you have to take into account that Lightning Network and other approaches will very probably mean that the on-chain volume will grow slower.

[ComputerGenie]

If dreams, wishes, and delusions earned an income, some of you folks would be rich. 

[dinofelis]

If dreams, wishes, and delusions earned an income, some of you folks would be rich. 

I thought that was the basic principle of crypto

[dinofelis]

On blockchain.info you can also find charts that indicate that the transaction rate grows less than 50% per year. That would mean that with Segwit alone blocks will be full at the end of 2018, and with Segwit2x you have time until 2020/21. But you have to take into account that Lightning Network and other approaches will very probably mean that the on-chain volume will grow slower.

I wouldn't touch LN with a stick on a hard-limited block chain where you have no guarantee of settlement.   The LN is an entirely different way of doing transactions that has never been experimented, nobody knows if *in practice* it will take off.  The step from on-chain transactions to LN is about as big as the change from, say, using a credit card to using a crypto.  Maybe it will be a success, maybe not at all.  It's something entirely different than what crypto was about until now.   The LN would be a great way to connect to one's exchange, that's true.  It would keep the exchanges honest, and at the same time, it would turn the exchanges into the banks of tomorrow.  But that would entirely change how people see crypto, namely essentially "back to banking". 
If an exchange has a serious problem, all its customers would need to settle immediately on chain ; if they can't, because the chain is too small, this is a way for the exchange to run with the money as usual.

[d5000]

I wouldn't touch LN with a stick on a hard-limited block chain where you have no guarantee of settlement.   The LN is an entirely different way of doing transactions that has never been experimented, nobody knows if *in practice* it will take off.

I also think it is a little bit overrated (for me it will never replace on-chain transactions entirely), but in principle it can be a great micropayments tool. That's also why my answer to ComputerGenie (who preferred to omit a response to my answer) was cautious, I don't say on-chain volume will go down after a successful LN implementation, but on-chain volume would grow a bit slower because LN would be the main tool for transactions of e.g. $5 or $20 and so a part of the pressure is taken from the "transactions market".

[dinofelis]

I wouldn't touch LN with a stick on a hard-limited block chain where you have no guarantee of settlement.   The LN is an entirely different way of doing transactions that has never been experimented, nobody knows if *in practice* it will take off.

I also think it is a little bit overrated (for me it will never replace on-chain transactions entirely), but in principle it can be a great micropayments tool. That's also why my answer to ComputerGenie (who preferred to omit a response to my answer) was cautious, I don't say on-chain volume will go down after a successful LN implementation, but on-chain volume would grow a bit slower because LN would be the main tool for transactions of e.g. $5 or $20 and so a part of the pressure is taken from the "transactions market".

Well, looking at blockchain.info, one can see that the number of daily transactions has been rising slowly until about 300 000 transactions per day, and that the average estimated daily transaction VOLUME is about 250 000 BTC daily since quite a while.  So we see that the *average* bitcoin transaction transacts about 1 BTC.  I have no idea of its distribution, but apparently a transaction on chain is around 1 BTC : are there 240 000 transactions of micro-amounts, and 10 000 transactions of several tens of BTC, or are all of them turning around 1 BTC, I don't know.  But the average transaction doesn't seem to be a micro transaction.
Of course, the bitcoin block chain is also used as an immutable ledger for other applications than transactions ; these will not go away.

[d5000]

Well, looking at blockchain.info, one can see that the number of daily transactions has been rising slowly until about 300 000 transactions per day, and that the average estimated daily transaction VOLUME is about 250 000 BTC daily since quite a while.

Yes, you're right here. It's even getting worse measured in USD: As Bitcoin has doubled in price since the end of 2016 the USD value of the average transaction has doubled.

But I'm sure that if we looked at the median, it would be much lower than the average (does anyone know the charts/numbers here?). I often observe a lot of transactions in the 0.01-0.1 BTC range. 0.1 would be a little high for LN, but the lower end (0.01-0.04, as LN has, if I remember right, a limitation to 0.042 BTC) could be covered with lightning transactions.

The current high average transaction amount may also to be related to the fact that businesses like exchanges may bundle a higher number of transactions together than they used to do before.

Of course, the bitcoin block chain is also used as an immutable ledger for other applications than transactions ; these will not go away.

Agree. But there is a large subset of the OP_RETURN transactions that could use LN (for example, Counterparty asset and other coloured coin transfers). A large number is using just the "immutable" feature provided by on-chain transactions (e.g. Factom), that's true.

[ComputerGenie]

... ComputerGenie (who preferred to omit a response to my answer)....

I did what, to who, about what? 

[d5000]

... ComputerGenie (who preferred to omit a response to my answer)....

I did what, to who, about what? 

Well, here I think I provided strong arguments to back my claim that it is very probable to almost sure that in the case of an introduction of Segwit users would pay less transaction fees (I know it's not a mathematical proof, but that's impossible if we're trying to predict human behaviour).

Your next post could be interpreted as an accusation to me to base my stance on "dreams, wishes, and delusions".

[ComputerGenie]

... ComputerGenie (who preferred to omit a response to my answer)....

I did what, to who, about what? 

Well, here I think I provided strong arguments to back my claim that it is very probable to almost sure that in the case of an introduction of Segwit users would pay less transaction fees (I know it's not a mathematical proof, but that's impossible if we're trying to predict human behaviour).

Your next post could be interpreted as an accusation to me to base my stance on "dreams, wishes, and delusions".

OH, you mean the post where you ignored that it would take 50 2M blocks just to clear the current mempool if no new transactions ever occur, that "everyone" claims there are transactions that don't occur because there's a backlog, and where "everyone" ignores the amount of transactions that they claim aren't happening because of congestion?