Transaction fee will rise, regardless of blocksize

[johnyj]

In a discussion about transaction fee, I just sorted out these facts:

A smart mining pool will calculate the block orphan risk when including transactions (more transactions will raise the orphan rate due to longer broadcasting time)

For example, if a full 1MB block containing 2000 transactions will broadcast slower than an empty block, thus raise the block orphan rate by 1%, then the pool will lose 1% of the block reward, which is 0.25 BTC, divided by 2000 transactions, you get 0.000125 btc per transaction, that's the recommended fee right now

For 2MB blocks, the broadcasting time of such a large block might affect the orphan rate more significantly, say raise the block orphan rate to 4%, and then you can include 4000 transactions in each block. As a result, although each block can hold double amount of transactions, a full block will cost 4x more for miners, so the fee for each transaction would double at 2 MB full block

Currently, due to the block reward, you can not enforce a higher fee for miners, since the miners on today's fast network have very low orphan rate if they include all the 2000 transactions, thus any fee above 0.0001 is good enough for them to happily include it

However, if block is larger and the transactions processing speed (CPU/network) is limited, miners can not include as many as possible transactions without added orphan cost, then I guess the fee will rise when blocks grow beyond certain size, regardless of the block size limit

Only a speed increase in CPU/network infrastructure would lower these cost, and we know that the improvement in those area (especially the fiber network infrastructure) are not that easy to upgrade

[VirosaGITS]

In a discussion about transaction fee, I just sorted out these facts:

A smart mining pool will calculate the block orphan risk when including transactions (more transactions will raise the orphan rate due to longer broadcasting time)

For example, if a full 1MB block containing 2000 transactions will broadcast slower than an empty block, thus raise the block orphan rate by 1%, then the pool will lose 1% of the block reward, which is 0.25 BTC, divided by 2000 transactions, you get 0.000125 btc per transaction, that's the recommended fee right now

For 2MB blocks, the broadcasting time of such a large block might affect the orphan rate more significantly, say raise the block orphan rate to 4%, and then you can include 4000 transactions in each block. As a result, although each block can hold double amount of transactions, a full block will cost 4x more for miners, so the fee for each transaction would double at 2 MB full block

Currently, due to the block reward, you can not enforce a higher fee for miners, since the miners on today's fast network have very low orphan rate if they include all the 2000 transactions, thus any fee above 0.0001 is good enough for them to happily include it

However, if block is larger and the transactions processing speed (CPU/network) is limited, miners can not include as many as possible transactions without added orphan cost, then I guess the fee will rise when blocks grow beyond certain size, regardless of the block size limit

Only a speed increase in CPU/network infrastructure would lower these cost, and we know that the improvement in those area (especially the fiber network infrastructure) are not that easy to upgrade

The block propagation still depend on location and the time for the block to fully propagate (and thus be safe) depend from where it start. So needing better hardware to compute it faster is one thing, but just getting better hardware will not fully solve that.

So the price would go up per tx as well, yes.

[Patel]

It SHOULD be expensive to have your transaction in the blockchain.

Offchain transactions should be used for majority of tx's. It just so happens that payment channels is a decentralized off-chain method.

[USB-S]

It SHOULD be expensive to have your transaction in the blockchain.

Offchain transactions should be used for majority of tx's. It just so happens that payment channels is a decentralized off-chain method.

There aren't any good usable off chains available at the moment. Getting the end you and end user use the same off chain may be troublesome as well.
Coinbase has his stuff going, but being outside of US renders coinbase useless for me. Haven't looked into other offchains though.

[VirosaGITS]

It SHOULD be expensive to have your transaction in the blockchain.

Offchain transactions should be used for majority of tx's. It just so happens that payment channels is a decentralized off-chain method.

There aren't any good usable off chains available at the moment. Getting the end you and end user use the same off chain may be troublesome as well.
Coinbase has his stuff going, but being outside of US renders coinbase useless for me. Haven't looked into other offchains though.

Not particularly salient but the service is also available in Canada for cashing in and out. US does have more features though, like the BTC debit card.

Regardless for the time being, including your transaction is pretty cheap so no biggie.

[NorrisK]

It SHOULD be expensive to have your transaction in the blockchain.

Offchain transactions should be used for majority of tx's. It just so happens that payment channels is a decentralized off-chain method.

Don't the offchain transaction require you to put trust in a thrid party again? That seems to me something that you really don't want to do if you are into bitcoin for the reason of decentralization..
Bitcoin was supposed to be cheap to send value around the world, and that is not compatible with increasing fees.

[VirosaGITS]

It SHOULD be expensive to have your transaction in the blockchain.

Offchain transactions should be used for majority of tx's. It just so happens that payment channels is a decentralized off-chain method.

Don't the offchain transaction require you to put trust in a thrid party again? That seems to me something that you really don't want to do if you are into bitcoin for the reason of decentralization..
Bitcoin was supposed to be cheap to send value around the world, and that is not compatible with increasing fees.

No, the way they want to do it, it would be a blockchain, just like the Bitcoin block chain, it would be decentralized and there would simply be code added in later core version to allow tx to leave one chain to enter the other.

I'm sure there could be 3rd party services trying to sell their sidechain too but i'm not sure how possible it will be for people to just integrate sidechains like that.

[yenxz]

I dont care if it rise,i think it not gonna big fee. And if it rise,i think just double or triple

[Jeremycoin]

It's okay for me if the transaction fee goes higher, as long as it's safe.
However, don't rise to much. Cuz it doesn't make sense if you want to send a $1 with $2 fee

[Amph]

no problem for that, because the vlaue will also increase, which mean that we will spend less bitcoin, which mean less fee as a rsult, so the thing will be levelled in the end

also they are still so cheap that even a big increase would still put them in the "cheap territory"

[n2004al]

In a discussion about transaction fee, I just sorted out these facts:

A smart mining pool will calculate the block orphan risk when including transactions (more transactions will raise the orphan rate due to longer broadcasting time)

For example, if a full 1MB block containing 2000 transactions will broadcast slower than an empty block, thus raise the block orphan rate by 1%, then the pool will lose 1% of the block reward, which is 0.25 BTC, divided by 2000 transactions, you get 0.000125 btc per transaction, that's the recommended fee right now

For 2MB blocks, the broadcasting time of such a large block might affect the orphan rate more significantly, say raise the block orphan rate to 4%, and then you can include 4000 transactions in each block. As a result, although each block can hold double amount of transactions, a full block will cost 4x more for miners, so the fee for each transaction would double at 2 MB full block

Currently, due to the block reward, you can not enforce a higher fee for miners, since the miners on today's fast network have very low orphan rate if they include all the 2000 transactions, thus any fee above 0.0001 is good enough for them to happily include it

However, if block is larger and the transactions processing speed (CPU/network) is limited, miners can not include as many as possible transactions without added orphan cost, then I guess the fee will rise when blocks grow beyond certain size, regardless of the block size limit

Only a speed increase in CPU/network infrastructure would lower these cost, and we know that the improvement in those area (especially the fiber network infrastructure) are not that easy to upgrade

The miners can include less transactions and add the same the cost of orphan blocks. Using the same fees. If so, I think that are not to many the transactions which will be less, (comparing to the situation with higher fees and the maximum of transactions) and this cannot delay to much the time of confirmations. If there were be more higher fees bitcoin will (risk to) lose one of the most its precious qualities. Its ability to compete with the other kind of transactions made with the normal money. So, even the above write by me cannot be possible or not preferable, must be found another solution which can make possible the not increase of the fees.

[bearex]

If it does, than Bitcoin kind of loses its meaning.

[TastyChillySauce00]

if its just 4 cents, it'll not be a big problem, im not that stingy to complaint about losing 1 cents, and nowadays 1 cent isn't worth of anything

[Tey]

In a discussion about transaction fee, I just sorted out these facts:

For example, if a full 1MB block containing 2000 transactions will broadcast slower than an empty block, thus raise the block orphan rate by 1%, then the pool will lose 1% of the block reward..

Please can you elaborate on how you concluded the 1% rate ?

[btckold24]

call me cheap but I hate fee's going up. Think about all the fee's you have paid if btc was at 10k per
which we hope it will be one day. On that note I also hate taxes lol. Maybe i'm just cheap....

[johnyj]

It SHOULD be expensive to have your transaction in the blockchain.

Offchain transactions should be used for majority of tx's. It just so happens that payment channels is a decentralized off-chain method.

Don't the offchain transaction require you to put trust in a thrid party again? That seems to me something that you really don't want to do if you are into bitcoin for the reason of decentralization..
Bitcoin was supposed to be cheap to send value around the world, and that is not compatible with increasing fees.

The offchain solution is the only way to compete with traditional payment capacity. As I'm familiar with, in some countries like China Sweden etc..., mobile payment and third party payment solutions do not charge any fee at all, and the payment is done instantly, so bitcoin blockchain has no way to catch up with those solutions, only centralized offchain solution can compete

But the bitcoin's main advantage is its limited supply and censorship resistant, not the payment function

[johnyj]

no problem for that, because the value will also increase, which mean that we will spend less bitcoin, which mean less fee as a result, so the thing will be leveled in the end

also they are still so cheap that even a big increase would still put them in the "cheap territory"

IMO, the block orphan rate affects the mining income of bitcoins, has no direct relation to bitcoin's price

If one bitcoin is worth 1 million dollars, and miners still get 1% orphan rate by including 2000 transactions (e.g. lose 0.25 BTC), so they would still charge 0.25 BTC for 2000 transactions, 0.000125 btc per transaction, but then each transaction will cost $125, anything below that will not be included by the miners, since they get more than that by not including such a transaction

In that case, if the network speed do not increase by 3-5 magnitudes, then transaction on the blockchain will become so expensive that small casual spending would only use off-chain-zero-fee services. Hopefully by the time bitcoin reach 1 million dollars, the block reward has dropped to single digits, and hardware and network would reach that level of capacity

[johnyj]

In a discussion about transaction fee, I just sorted out these facts:

For example, if a full 1MB block containing 2000 transactions will broadcast slower than an empty block, thus raise the block orphan rate by 1%, then the pool will lose 1% of the block reward..

Please can you elaborate on how you concluded the 1% rate ?

Actually 1%, e.g. 6 seconds delay in 600 seconds block interval is extremely low, in practice all the direct measurement value is much higher than this, see the paper below

https://scalingbitcoin.org/papers/feemarket.pdf
---------------------------------------------------------------------------------------------------------------------------
Table 2.  Estimates for the Block Solution Propagation Impedance and the Associated Minimum Fee Density

Estimate name                Propagation impedance(sec/MB)  Method                         Year   Ref.    Min. fee density (μɃ/kB)

Decker and Wattenhofer   80                                            Direct measurement       2013  10      3,333
Tradeblock                      17.1                                          Direct measurement       2015  18      713
This study                       7.6                                           Theoretically formula       2015  19      318
Relay network                 2.0                                           Taken as T5 time to 99%  2015  20     83
---------------------------------------------------------------------------------------------------------------------------

I would take the conservative approach of 60 seconds delay, about 10% of the block interval, which means current fee is not optimal from miners point of view. Obviously miners are still mainly incentivized by the block reward, so they would sacrifice some of the fee income to maintain the health of the network. But those miners on slower network (for example chinese pools) might start to generate empty blocks to improve their propagation speed: If a large mining pool are mining 10 blocks per day, losing 1 block of that income is a huge loss

[CIYAM]

But those miners on slower network (for example chinese pools) might start to generate empty blocks to improve their propagation speed

That would of course not be good for the network (and we have seen single tx blocks before).

This is why trying to *force* bigger blocks on miners (especially those in China) could backfire very badly (the result could end up being blocks with only the coinbase tx in them rather than the many MBs of txs people are expecting to see as there is simply no consensus rule that requires a miner to include any tx other than the coinbase one which itself is already likely optional because eventually there will be no block reward).

[countryfree]

I don't see anything wrong with an increase in transaction fees. If it comes with more safety, and more speed, I'll favor it. Frankly, as long as the fee remains below 10 cents, I don't see why anyone in the first world should care. It's only normal that we pay for such a great service.