BIP - Block75 - Historical and future projections (t. khan)
3 messages, 10 Jan 2017. Started by Ryan J Martin.
Ryan J Martin, Adam Back, t. khan
The adaptive/automatic block size notion has been around for a while--- others would be better able to speak to why it hasn't gotten traction. However my concern with something like that is that it doesn't regard the optimal economic equilibrium for tx fees/size---not that the current limit does either but the concern with an auto-adjusting size limit that ignores this would be the potential to create unforeseen externalities for miners/users. Miners may decide it is more profitable to mine very small blocks to constrict supply and increase marginal fees and with how centralized mining is, where a dozen pools have 85% hashrate, a couple of pools could do this. Then on the other side, maybe the prisoner's dilemma would hold and all miners would have minrelaytxfee set at zero and users would push the blocks to larger and larger sizes causing higher and higher latency and network issues.
Perhaps something like this could work (I can only speak to the economic side anyway) but it would have to have some solid code that has a social benefit model built in to adjust to an equilibrium that is able to optimize---as in maximizes benefit/minimize cost for both sides via a Marshallian surplus model--- for each size point.
To be clear, I'm not saying an auto-adjusting limit is unworkable (again only from an economic standpoint), just that it would need to have these considerations built in.
-Ryan J. Martin
________________________________________
------------------------------
Message: 2
Date: Mon, 9 Jan 2017 14:52:31 -0500
From: "t. khan" <teekhan42@gmail•com>
To: Bitcoin Protocol Discussion
<bitcoin-dev@lists•linuxfoundation•org>
Subject: [bitcoin-dev] BIP - Block75 - Historical and future
projections
Message-ID:
<CAGCNRJpSV9zKxhVvqpMVPyFyXco_ABB9a7_ihaDKEKFPQ9v3sw@mail•gmail•com>
Content-Type: text/plain; charset="utf-8"
Using daily average block size over the past year (source: https://blockchain.info/charts/avg-block-size?daysAverageString=14×pan=1year ), here's how Block75 would have altered max block sizes:
[image: Inline image 1]
As of today, the max block size would be 1,135KB.
Looking forward and using the last year's growth rate as a model:
[image: Inline image 2]
This shows the max block size one year from now would be 2,064KB, if Block75 activated today.
Of course, this is just an estimate, but even accounting for a substantial increase in transactions in the last quarter of 2017 and changes brought about by SegWit (hopefully) activating, Block75 alters the max size in such a way that allows for growth, keeps blocks as small as possible, *and* maintains transaction fees at a level similar to May/June 2016.
If anyone has an alternate way to model future behavior, please run it through the Block75 algorithm.
Every 2016 blocks, do this:
new max blocksize = x + (x * (AVERAGE_CAPACITY - TARGET_CAPACITY))
TARGET_CAPACITY = 0.75 //Block75's target of keeping blocks 75% full AVERAGE_CAPACITY = average percentage full of the last 2016 blocks, as a decimal x = current max block size
Thanks,
- t.k. -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20170109/b0e0b713/attachment.html> -------------- next part -------------- A non-text attachment was scrubbed... Name: Block75 2016.png Type: image/png Size: 32088 bytes Desc: not available URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20170109/b0e0b713/attachment.png> -------------- next part -------------- A non-text attachment was scrubbed... Name: Block75 2017.png Type: image/png Size: 33176 bytes Desc: not available URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20170109/b0e0b713/attachment-0001.png>
See discussion on bitcoin-discuss on this topic last few days. People may want to subscribe to that for more free wheeling discussion.
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-discuss
Adam
On 10 January 2017 at 04:14, Ryan J Martin via bitcoin-dev <bitcoin-dev@lists•linuxfoundation•org> wrote:
Quoted text (91 lines)
> The adaptive/automatic block size notion has been around for a while--- others would be better able to speak to why it hasn't gotten traction. However my concern with something like that is that it doesn't regard the optimal economic equilibrium for tx fees/size---not that the current limit does either but the concern with an auto-adjusting size limit that ignores this would be the potential to create unforeseen externalities for miners/users. Miners may decide it is more profitable to mine very small blocks to constrict supply and increase marginal fees and with how centralized mining is, where a dozen pools have 85% hashrate, a couple of pools could do this. Then on the other side, maybe the prisoner's dilemma would hold and all miners would have minrelaytxfee set at zero and users would push the blocks to larger and larger sizes causing higher and higher latency and network issues. > Perhaps something like this could work (I can only speak to the economic side anyway) but it would have to have some solid code that has a social benefit model built in to adjust to an equilibrium that is able to optimize---as in maximizes benefit/minimize cost for both sides via a Marshallian surplus model--- for each size point. > To be clear, I'm not saying an auto-adjusting limit is unworkable (again only from an economic standpoint), just that it would need to have these considerations built in. > > -Ryan J. Martin > > > ________________________________________ > > ------------------------------ > > Message: 2 > Date: Mon, 9 Jan 2017 14:52:31 -0500 > From: "t. khan" <teekhan42@gmail•com> > To: Bitcoin Protocol Discussion > <bitcoin-dev@lists•linuxfoundation•org> > Subject: [bitcoin-dev] BIP - Block75 - Historical and future > projections > Message-ID: > <CAGCNRJpSV9zKxhVvqpMVPyFyXco_ABB9a7_ihaDKEKFPQ9v3sw@mail•gmail•com> > Content-Type: text/plain; charset="utf-8" > > Using daily average block size over the past year (source: > https://blockchain.info/charts/avg-block-size?daysAverageString=14×pan=1year > ), here's how Block75 would have altered max block sizes: > > [image: Inline image 1] > > As of today, the max block size would be 1,135KB. > > Looking forward and using the last year's growth rate as a model: > > [image: Inline image 2] > > This shows the max block size one year from now would be 2,064KB, if > Block75 activated today. > > Of course, this is just an estimate, but even accounting for a substantial > increase in transactions in the last quarter of 2017 and changes brought > about by SegWit (hopefully) activating, Block75 alters the max size in such > a way that allows for growth, keeps blocks as small as possible, *and* > maintains transaction fees at a level similar to May/June 2016. > > If anyone has an alternate way to model future behavior, please run it > through the Block75 algorithm. > > Every 2016 blocks, do this: > > new max blocksize = x + (x * (AVERAGE_CAPACITY - TARGET_CAPACITY)) > > TARGET_CAPACITY = 0.75 //Block75's target of keeping blocks 75% full > AVERAGE_CAPACITY = average percentage full of the last 2016 blocks, as a > decimal > x = current max block size > > > Thanks, > > - t.k. > -------------- next part -------------- > An HTML attachment was scrubbed... > URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20170109/b0e0b713/attachment.html> > -------------- next part -------------- > A non-text attachment was scrubbed... > Name: Block75 2016.png > Type: image/png > Size: 32088 bytes > Desc: not available > URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20170109/b0e0b713/attachment.png> > -------------- next part -------------- > A non-text attachment was scrubbed... > Name: Block75 2017.png > Type: image/png > Size: 33176 bytes > Desc: not available > URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/attachments/20170109/b0e0b713/attachment-0001.png> > > ------------------------------ > > _______________________________________________ > bitcoin-dev mailing list > bitcoin-dev@lists•linuxfoundation•org > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev > > > End of bitcoin-dev Digest, Vol 20, Issue 21 > ******************************************* > _______________________________________________ > bitcoin-dev mailing list > bitcoin-dev@lists•linuxfoundation•org > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
As Block75 maintains blocks at 75% full (average over time), it automatically stabilizes transaction fees at about the level they were in May/June 2016. It will even do so through changes in transaction size and volume caused by SegWit and Lightning.
- t.k.
On Mon, Jan 9, 2017 at 11:14 PM, Ryan J Martin via bitcoin-dev < bitcoin-dev@lists•linuxfoundation•org> wrote:
Quoted text (115 lines)
> The adaptive/automatic block size notion has been around for a > while--- others would be better able to speak to why it hasn't gotten > traction. However my concern with something like that is that it doesn't > regard the optimal economic equilibrium for tx fees/size---not that the > current limit does either but the concern with an auto-adjusting size limit > that ignores this would be the potential to create unforeseen > externalities for miners/users. Miners may decide it is more profitable to > mine very small blocks to constrict supply and increase marginal fees and > with how centralized mining is, where a dozen pools have 85% hashrate, a > couple of pools could do this. Then on the other side, maybe the prisoner's > dilemma would hold and all miners would have minrelaytxfee set at zero and > users would push the blocks to larger and larger sizes causing higher and > higher latency and network issues. > Perhaps something like this could work (I can only speak to the > economic side anyway) but it would have to have some solid code that has a > social benefit model built in to adjust to an equilibrium that is able to > optimize---as in maximizes benefit/minimize cost for both sides via a > Marshallian surplus model--- for each size point. > To be clear, I'm not saying an auto-adjusting limit is unworkable > (again only from an economic standpoint), just that it would need to have > these considerations built in. > > -Ryan J. Martin > > > ________________________________________ > > ------------------------------ > > Message: 2 > Date: Mon, 9 Jan 2017 14:52:31 -0500 > From: "t. khan" <teekhan42@gmail•com> > To: Bitcoin Protocol Discussion > <bitcoin-dev@lists•linuxfoundation•org> > Subject: [bitcoin-dev] BIP - Block75 - Historical and future > projections > Message-ID: > <CAGCNRJpSV9zKxhVvqpMVPyFyXco_ABB9a7_ihaDKEKFPQ9v3sw@mail. > gmail.com> > Content-Type: text/plain; charset="utf-8" > > Using daily average block size over the past year (source: > https://blockchain.info/charts/avg-block-size? > daysAverageString=14×pan=1year > ), here's how Block75 would have altered max block sizes: > > [image: Inline image 1] > > As of today, the max block size would be 1,135KB. > > Looking forward and using the last year's growth rate as a model: > > [image: Inline image 2] > > This shows the max block size one year from now would be 2,064KB, if > Block75 activated today. > > Of course, this is just an estimate, but even accounting for a substantial > increase in transactions in the last quarter of 2017 and changes brought > about by SegWit (hopefully) activating, Block75 alters the max size in such > a way that allows for growth, keeps blocks as small as possible, *and* > maintains transaction fees at a level similar to May/June 2016. > > If anyone has an alternate way to model future behavior, please run it > through the Block75 algorithm. > > Every 2016 blocks, do this: > > new max blocksize = x + (x * (AVERAGE_CAPACITY - TARGET_CAPACITY)) > > TARGET_CAPACITY = 0.75 //Block75's target of keeping blocks 75% full > AVERAGE_CAPACITY = average percentage full of the last 2016 blocks, as a > decimal > x = current max block size > > > Thanks, > > - t.k. > -------------- next part -------------- > An HTML attachment was scrubbed... > URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/ > attachments/20170109/b0e0b713/attachment.html> > -------------- next part -------------- > A non-text attachment was scrubbed... > Name: Block75 2016.png > Type: image/png > Size: 32088 bytes > Desc: not available > URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/ > attachments/20170109/b0e0b713/attachment.png> > -------------- next part -------------- > A non-text attachment was scrubbed... > Name: Block75 2017.png > Type: image/png > Size: 33176 bytes > Desc: not available > URL: <http://lists.linuxfoundation.org/pipermail/bitcoin-dev/ > attachments/20170109/b0e0b713/attachment-0001.png> > > ------------------------------ > > _______________________________________________ > bitcoin-dev mailing list > bitcoin-dev@lists•linuxfoundation•org > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev > > > End of bitcoin-dev Digest, Vol 20, Issue 21 > ******************************************* > _______________________________________________ > bitcoin-dev mailing list > bitcoin-dev@lists•linuxfoundation•org > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev >