Gold collapsing. Bitcoin UP.

cypherdoc

Well-Known Member
Aug 26, 2015
5,257
12,998

[doublepost=1454520522,1454519799][/doublepost]

So the correct answer is to allow the market (miners) to decide rather than hard-code it in with a huge discount for segwit transactions.
so you agree with my assessment of the math and it's economic consequences? i mean, you can see how the more LN multi sigs get added to a block, the less space regular tx's get and the more miners are forced to accept discounted fees?
[doublepost=1454521643][/doublepost]well, what did i say? the mkt is anticipating further dollar devaluation. big move:

 
  • Like
Reactions: AdrianX

YarkoL

Active Member
Dec 18, 2015
176
258
Tuusula
yarkol.github.io
@cypherdoc
The 4MB is a "worst case" according to Maxwell.. of course it is possible that there is a cunning plan to make that a norm, but I would be at this stage cautious to state things like that. As for the discounted LN traffic, I guess a case could be made that the miners are compensated by the bigger fees due to 1MB limit and artificial fee market.
I'm a long way from wrapping all this around my head, and one would need to have an in-depth analysis of the off-chain transactions vs. classic transactions and their economic and technical implications.
 

tynwald

Member
Dec 8, 2015
69
176
c'mon guys. no feedback on my technical post above from Towns?
<snip>
get your heads around the math and give me feedback.
cypherdoc, great work and I think the results show one very strong reason why SW is being pushed so hard, so fast by Blockstream.

2 suggestions:

1) setup a spreadsheet on google docs etc to show the effect on miner income under different ratio of transaction types.
2) translate your argument into chinese and post to the sites read by chinese bitcoin fans, as well sending to chinese mining pools.

Money (loss of) motivates people faster than words on a web page.
 

sgbett

Active Member
Aug 25, 2015
216
786
UK
c'mon guys. no feedback on my technical post above from Towns?

this is incredibly important b/c heretofore, everyone likes SW. there has been NO criticism of it either from a technical or economic standpoint. if i'm right, we now have ammunition to spread to the rest of the community as to why SW will be bad for mainchain scaling and miner long term health. this could sway the discussion.

get your heads around the math and give me feedback.
"Yeah its fucked up" is my feedback on it.

I'm sad and confused, the same pattern keeps playing out. I don't want to have to be constantly cynical and untrusting. Someone comes up with segwit and pitches the benefits of it and being a reasonable person you assume that everyone else is reasonable, and that segwit is good! You want to see the good in things. Then it all turns to dust. Again.

Its annoying because taking each individual thing you can't conclusively say "yes there is this blatant corruption thing going on" I still want to believe in the inherent good nature of (most) people, but one thing after another keeps cropping up.

When you put all these things to gather it makes the baby jesus cry :/
 

cypherdoc

Well-Known Member
Aug 26, 2015
5,257
12,998
@YarkoL

let me repost what AJTowns said:

The 4MB consensus limit could only be hit by having a single trivial
transaction using as little base data as possible, then a single huge
4MB witness. So people trying to abuse the system have 4x the blocksize
for 1 block's worth of fees, while people using it as intended only get
1.6x or 2x the blocksize... That seems kinda backwards.


why do you think he says it's backwards?
 
  • Like
Reactions: AdrianX

Mengerian

Moderator
Staff member
Aug 29, 2015
536
2,597
get your heads around the math and give me feedback.
SW adds a new consensus rule: a + b/4 <= 1MB, where a is the size of the "old" block, and b is the size of the "witness" (ie, the signatures). So yes, all else being equal, this consensus rule will bias miners to select transactions with small "base" components, but with lots of signature data. Of course, transaction creators can compensate by increasing the fees for transactions with simple signatures, so it's not obvious what kinds of transactions would end up being included in blocks.

The ostensible reason for all this is twofold 1) penalize transaction that can add to the size of the Unspent Transaction Output database 2) future plans to prune off signature data for downloading historical transactions and serving transactions to SPV wallets should mean that the signature portion will theoretically consume less bandwidth in the future.

These are some of the problems I see with this new "a + b/4" consensus rule:

- It is complicated, and adds technical debt. Ironic considering this comment by GMax:
(ht @albin )

- It is completely arbitrary. Why a 1/4 weighting? Why not 1/3, 1/6, 1/10?

- It penalizes transactions that reduce the size of the UTXO set the same as those that add to it. So it's not obvious it incentivises reducing the UTXO database.

- Changing the UTXO set is just another way of saying "updating the ledger". And as @Justus Ranvier pointed out, ease with which one can update the ledger is one of the key properties of Bitcoin that adds to its value. So restricting changes to the UTXO set is bad for the value of the Bitcoin network.
 
Last edited:

cypherdoc

Well-Known Member
Aug 26, 2015
5,257
12,998
@Mengerian

i'm not following. "a" in Towns example is the size of "regular tx's" (not multi sig) dedicated to a block. i think.

it also seems to me that miners would rather incl these reg tx's in a block over the multi sigs b/c they will pay 4x the fee and involve transmitting less data.

edited: that can't be right. "a" must be both reg & multi-sigs.
 
Last edited:

Mengerian

Moderator
Staff member
Aug 29, 2015
536
2,597
@cypherdoc
Yes, "regular" transactions would add solely to "a". SW transactions add to both "a" and "b".

Which transactions miners prefer to include might depend partially on minimizing data transmission. But it could also depend on other constraints, like how many transactions they can pack into the block-size limit (which would favor transactions with more complex signatures, for larger b), or how long they take to validate (which favors transactions with less signature data, for smaller b).
 
  • Like
Reactions: cypherdoc

cypherdoc

Well-Known Member
Aug 26, 2015
5,257
12,998
@Mengerian et al

ok guys, then i need to re-evaluate much of what i said above regarding AJTowns calcs.
 

solex

Moderator
Staff member
Aug 22, 2015
1,558
4,694
c'mon guys. no feedback on my technical post above from Towns?

this is incredibly important b/c heretofore, everyone likes SW. there has been NO criticism of it either from a technical or economic standpoint. if i'm right, we now have ammunition to spread to the rest of the community as to why SW will be bad for mainchain scaling and miner long term health. this could sway the discussion.

get your heads around the math and give me feedback.
@cypherdoc
Your discussion with Antony Towns is very interesting and revealing, thanks for posting it. I have a lot of respect for Towns' ability to dig into blockchain math.

I was initially enthusiastic about SW after Pieter's presentation, and like many people first took that 4MB at face value, and it is only all the debate since that has tempered it with more knowledge about the pros & cons. It still has huge pros, such as the fraud-proofs, though BS-Core are leaving this on the back-burner to focus on what BS needs first.

The strategy for non-Core Bitcoiners is to deal with 1 fight at a time, and getting the 2MB hard-fork done, immediately puts the whole SW initiative into true context, an enhancement standing on its own merits without being forced on the community as a partial solution to the current full-blocks problem.

After the 2MB is done then SW will have to be argued by the rump of Core to the Classic, BU & XT user base. Maybe it will be like RBF and go down like a lead balloon. Maybe it can be modified so that LN and others pay the same per byte for their massive multi-sigs as ordinary simple tx. I think the latter is a minimum.

tl;dr; Win the 2MB then SW needs to stand completely on its own technical and economic merit.
 
Last edited:

AdrianX

Well-Known Member
Aug 28, 2015
2,097
5,797
bitco.in
"Yeah its fucked up" is my feedback on it.

I'm sad and confused, the same pattern keeps playing out. I don't want to have to be constantly cynical and untrusting. Someone comes up with segwit and pitches the benefits of it and being a reasonable person you assume that everyone else is reasonable, and that segwit is good! You want to see the good in things. Then it all turns to dust. Again.

Its annoying because taking each individual thing you can't conclusively say "yes there is this blatant corruption thing going on" I still want to believe in the inherent good nature of (most) people, but one thing after another keeps cropping up.

When you put all these things to gather it makes the baby jesus cry :/
It's simple, let them do what they must, on condition they give up control of Bitcoin development, by all means fork Core to Blockstream Core, but let Bitcoin scale organically as designed.

I'll run Blockstream Core in a heartbeat if Satoshi's Bitcoin vision is unsustainable.

tl;dr; Win the 2MB then SW needs to stand completely on its own technical and economic merit.
@solex I think it's entirely possible if we win the 2MB block we will inevitably move on to the need for a 4MB block before we have need for more creative hacking of the blockchain.
 
Last edited:

sickpig

Active Member
Aug 28, 2015
926
2,541

BldSwtTrs

Active Member
Sep 10, 2015
196
583
Why Coinbase didn't hire some core devs? If they did so, the story would be different today.

Just pay the damn devs and watch their ideas conform to their economic interest. It's human nature.

Why nobody in the Bitcoin industry, except Blockstream, have had this idea until now?
 

cypherdoc

Well-Known Member
Aug 26, 2015
5,257
12,998
a dark day indeed but still holding:


[doublepost=1454545947,1454544965][/doublepost]ok, so the Chinese have been devaluing. then the Japanese massively devalued on Friday. the market is telling you the Fed will not be out done. after all, they like to stay in step. this should be good for fixed supply currencies like Bitcoin despite today's bad news:

 

Richy_T

Well-Known Member
Dec 27, 2015
1,085
2,741
so you agree with my assessment of the math and it's economic consequences? i mean, you can see how the more LN multi sigs get added to a block, the less space regular tx's get and the more miners are forced to accept discounted fees?
Absolutely. If the miners don't patch this, they're stupid. Assuming they're not, that makes segwit in an inconvenient place for Core.
 

Richy_T

Well-Known Member
Dec 27, 2015
1,085
2,741
It should be noted that the segwit subsidy is not a consensus rule and thus miners are free to discard it.
[doublepost=1454550533][/doublepost]
@cypherdoc
Yes, "regular" transactions would add solely to "a". SW transactions add to both "a" and "b".

Which transactions miners prefer to include might depend partially on minimizing data transmission. But it could also depend on other constraints, like how many transactions they can pack into the block-size limit (which would favor transactions with more complex signatures, for larger b), or how long they take to validate (which favors transactions with less signature data, for smaller b).
Miners likely prefer transactions which pay more. Factors like network transmission and storage are likely to be marginal considerations.

The point is though that these calculations should be left to miners and not... well, whatever shenanigans are going on here.