Richy_T
Well-Known Member
- Dec 27, 2015
- 1,085
- 2,741
Gold collapsing. Bitcoin UP.
- Thread starter cypherdoc
- Start date
Dusty
Active Member
- Mar 14, 2016
- 362
- 1,172
That would not be wise: while the block size would be the same, the consensus algorithm is not the same: classic has a limit on the maximum size of a transaction (if I remember correctly) and on the maximum number of byte hashed allowed in a tx.
This means that an attacker could create a special crafted transaction that would fork the network between the users of Classic and the "2mb patched core".
@Dusty:
The fork split you refer to could only occur if the majority of hash power was using the "2MB patched core" as well. Since this is highly unlikely, IMO it is completely safe for non-mining nodes to run the 1-byte patch.
And in the extremely unlikely event that a fork occurred (inconceivable IMO), the Classic nodes would eventually switch to the longest chain manually. Another reason that they should be running a Bitcoin-Unlimited-esque consensus tracking algorithm so that this happens automatically.
The fork split you refer to could only occur if the majority of hash power was using the "2MB patched core" as well. Since this is highly unlikely, IMO it is completely safe for non-mining nodes to run the 1-byte patch.
And in the extremely unlikely event that a fork occurred (inconceivable IMO), the Classic nodes would eventually switch to the longest chain manually. Another reason that they should be running a Bitcoin-Unlimited-esque consensus tracking algorithm so that this happens automatically.
@satoshis_sockpuppet
"If you give them an inch they will take a mile."
Now who the hell does that sound like? For years now kore dev has been running rough shod over the code with SF's.
No more!
[doublepost=1465161576][/doublepost]There's an optional 100kB tx limit that is off by default. Sigops limit is 1.3GB irc.
"If you give them an inch they will take a mile."
Now who the hell does that sound like? For years now kore dev has been running rough shod over the code with SF's.
No more!
[doublepost=1465161576][/doublepost]There's an optional 100kB tx limit that is off by default. Sigops limit is 1.3GB irc.
Okaaay...
[doublepost=1465162040][/doublepost]I should probably leave this to competent people...the point is to allow larger inbound blocks directly, without the 75% and grace period fluff.
Well my maths holds then...
Please can you actually think this through. Do you understand the massive difference with respect to 51% and a hardfork versus a double spend? This is a mathematical factual difference, not a political opinion.
For example look at section 11 of the Bitcoin whitepaper. It has calculations of an "attacker" trying to generate a longer chain than the "honest chain". It says:
Source: Bitcoin Whitepaper section 11
Now once again, the 1MB rule might be stupid, temporary, crippling Bitcoin or whatever negative you want to use. However, it is still an actual rule and you need to appreciate and understand how to remove it and that it is a fundamentally different process to a double spend with Nakamorto consensus.
The paper then goes on to give statistical examples, it says if an attacker has 30% of the hash-rate and is 10 blocks behind the honest chain, there is only 4.17% chance the attacker has of catching up.
These probabilities are entirely different before imposing rule changes, this maths does not apply. It is merely a different thing.
You keep framing this as a nefarious process. It is a market based process, which involves:
- Miners flagging support
- Nodes flagging support
- Coin votes
- And yes discussions and meetings
Now instead of carrying on doing the same, lets look at why you keep being defeated. Why is Peter Todd consistently able to run circles around you? It is because you do not appear to understand these processes and how these work, you do not understand there is a natural bias in favor of the status quo and to overcome this you have to be pretty decisive and ruthless. The Core team on the other hand have a very strong knowledge of this process. It may not be easy for you to hear this, but its true.
There is a strong bias in favor of staying on one chain, that is how the system was designed. Therefore changing the rules is a higher bar than Nakamoto consensus, you need the nodes on your side, even with a the majority of miners it is statistically harder, miners have a natural fear of splitting the chain or losing as they do not want to lose money.
Each time you are defeated to Core their reputation and power increases and yours diminishes. Core has the momentum.
Once you understand that you need strong momentum, you need a resounding and decisive victory, you need to demonstrate strength to overcome the inherent designed bias in favor of the status quo. You can not have a narrow or marginal victory, the miners and economic majority will not let you. Once you learn this, you will win.
Last edited:
"You can not have a narrow or marginal victory, the miners and economic majority will not let you. Once you learn this, you will win." -- @jonny1000
Don't think I agree, but it does sound very Yoda.
Don't think I agree, but it does sound very Yoda.
Last edited:
Richy_T
Well-Known Member
- Dec 27, 2015
- 1,085
- 2,741
Now, hold on a second.
Earlier, you said even if you had 75% of the hash rate and were making chains with bigger blocks, there was a 20% chance that the smaller chain could overtake and invalidate your blocks.
25<30 and 20>4.17
Putting aside loaded terms such as "attacker", these things do not agree with each other.
[doublepost=1465172419][/doublepost]
Hah, that old tactic, "Declare victory and go home". I've got some bad news for you, sunshine. The fight is only beginning. What you've seen from the "change the limit" side has just been trying to avoid the shit hitting the fan.
[doublepost=1465172736][/doublepost]
This is how RBF slipped in with hardly a whisper despite some quite solid opposition from the community. Because it is part of the status quo. Oh, wait.
Core has no understanding of these processes. They just have their hand on the tiller and the bully pulpit.
Last edited:
Satoshi's term
Well Satoshi's number assumes a 10 block lead already and then catching up to be level, my number assumes a 1 block lead and then overtaking to win.
How many times do I need to repeat, RBF has nothing to do with the rules for determining if blocks are valid. RBF is just some local mempool policy. RBF is neither a hardfrok or a softfork.
[doublepost=1465174143][/doublepost]
Well whether its necessary or not, why do you want such a narrow victory? What do you gain from it?
Richy_T
Well-Known Member
- Dec 27, 2015
- 1,085
- 2,741
He was using it in a different way to how you are. Your usage is causing you to make erroneous assumptions so you should probably consider stopping if you seek to make an honest argument.
OK, so just to be clear, you are taking directly comparable situations and using different goalposts to suit the argument you are trying to make.
Mathematically, there is no "attacker" which is purely being used as an appeal to emotion, there are only two chains of different hash power, one of which will accept the other chain as valid should it manage to become longer and the other one that wouldn't.
Or to be more precise, one group will accept the other chain and both groups will accept the chain with the old rules, giving the chain that does not change the rules a huge advantage, which is difficult to overcome, unless you are smart about it.
Please let me know if you understand this?
Richy_T
Well-Known Member
- Dec 27, 2015
- 1,085
- 2,741
Not that difficult. But it does have to be accounted for.
In truth, the odds are fairly small that there would be an issue soon after a >1MB block was produced (which would have little effect as it would most likely just mean a reordering of transactions) and would rapidly diminish as the chain with the larger hash power pulled ahead. Network affects would likely mean the smaller hash-power chain rapidly dwindling too.
Of course, things are not helped by Core having dragged things out so that transactions are often not timely so one which confirmed in a >1MB block which is in a chain that is invalidated could potentially take a long time to confirm on the crippled chain.
In truth, the odds are fairly small that there would be an issue soon after a >1MB block was produced (which would have little effect as it would most likely just mean a reordering of transactions) and would rapidly diminish as the chain with the larger hash power pulled ahead. Network affects would likely mean the smaller hash-power chain rapidly dwindling too.
Of course, things are not helped by Core having dragged things out so that transactions are often not timely so one which confirmed in a >1MB block which is in a chain that is invalidated could potentially take a long time to confirm on the crippled chain.
No, I was not. If you want the comparison, I think it is as follows:
For an "attacker" (Satoshi's term in Satoshi's context and in quotes) winning when they are one block behind and controls 30% of the hashrate:
- Normal double spend: c13% chance
- Doing a hardfork (when a better term than "attacker" could be used) : c0% chance
- Normal double spend: c43% chance
- Doing a hardfork: c0% chance
- Normal double spend: c100% chance
- Doing a hardfork: c0% chance
- Normal double spend: c100% chance
- Doing a hardfork: c50% chance
- Normal double spend: c100% chance
- Doing a hardfork: c79% chance
Last edited:
Richy_T
Well-Known Member
- Dec 27, 2015
- 1,085
- 2,741
Also consider that even if the crippled chain did get ahead somehow, there would be a backlog of transactions which would quickly lead to another >1MB block on the higher-limit chain which would then start to be built on again and the crippled chain would once again begin to fall behind. Any transactional anomalies would be confined to the period of a few hours and resolved with no loss of funds.
[doublepost=1465176371][/doublepost]We are not talking of double spends. I think you yourself said as much earlier.
You are also guilty of logical errors. A man attacking a castle wishes to enter. It does not follow that a man who wishes to enter a castle is an attacker.
An attacker is one who would be performing an action with the intent of detriment to the Bitcoin network and could expect to be opposed by those who would be harmed by the attack. A 2MB hard-fork could be seen as being net beneficial to users of the network and could be expected to be joined by those who share that viewpoint.
Your continued use of the term is really making it hard to accept you are acting in good faith as you claim.
[doublepost=1465176371][/doublepost]We are not talking of double spends. I think you yourself said as much earlier.
You are also guilty of logical errors. A man attacking a castle wishes to enter. It does not follow that a man who wishes to enter a castle is an attacker.
An attacker is one who would be performing an action with the intent of detriment to the Bitcoin network and could expect to be opposed by those who would be harmed by the attack. A 2MB hard-fork could be seen as being net beneficial to users of the network and could be expected to be joined by those who share that viewpoint.
Your continued use of the term is really making it hard to accept you are acting in good faith as you claim.
Last edited:
albin
Active Member
- Nov 8, 2015
- 931
- 4,008
I think we're long past the point of good faith assumption!
Now it's just on the level of douchey sematic fillibustering and snide patronizing.
This is obviously trying to give the lurkers some kind of twisted view of what's going on here, because he runs away just like Maxwell whenever anything challenging comes up.
Now it's just on the level of douchey sematic fillibustering and snide patronizing.
This is obviously trying to give the lurkers some kind of twisted view of what's going on here, because he runs away just like Maxwell whenever anything challenging comes up.
Richy_T
Well-Known Member
- Dec 27, 2015
- 1,085
- 2,741
Jonny, could you link to that Satoshi thing? I could probably find it but if you have it open already... I'd like to see the context.
[doublepost=1465177466][/doublepost]If you have 25% of the hashrate, your chance of finding the next block is 25% (in the case of the crippled chain, you are building on top of the last <1MB block, of course). Your chance of finding the next two blocks (and thus orphaning the >1MB block) would be 25% of 25% or ~6%. Congrats, you orphaned a block (nobody cares). Now you have to do it again.
If you are two blocks behind, you have to find the next three blocks. 25%*25%*25% or ~1.5% Could happen but still, nobody cares much.
Get towards invalidating an hours worth of transactions and it might cause some consternation. that's somewhere around .02%. And you probably mined most of those transactions yourself anyway (since this isn't actually an attack, transactions on both chains will be identical in nearly all cases just mined in different blocks).
You have a very small window to do this in, by the way. By the time the larger block chain is 10 blocks ahead it's pretty much game over. Meanwhile you're mining a chain that looks like a dead end. You're mining for the money, right? Time to switch.
[doublepost=1465177466][/doublepost]If you have 25% of the hashrate, your chance of finding the next block is 25% (in the case of the crippled chain, you are building on top of the last <1MB block, of course). Your chance of finding the next two blocks (and thus orphaning the >1MB block) would be 25% of 25% or ~6%. Congrats, you orphaned a block (nobody cares). Now you have to do it again.
If you are two blocks behind, you have to find the next three blocks. 25%*25%*25% or ~1.5% Could happen but still, nobody cares much.
Get towards invalidating an hours worth of transactions and it might cause some consternation. that's somewhere around .02%. And you probably mined most of those transactions yourself anyway (since this isn't actually an attack, transactions on both chains will be identical in nearly all cases just mined in different blocks).
You have a very small window to do this in, by the way. By the time the larger block chain is 10 blocks ahead it's pretty much game over. Meanwhile you're mining a chain that looks like a dead end. You're mining for the money, right? Time to switch.
Last edited:
You mean the calculation sector of the Bitcoin whitepaper?
Yes in the scenario the Classic chain has 75% of the miners and a 1 block lead, the chances of Classic winning are c82%. The chances of the less than 1MB chain winning in two blocks is 25% * 25% = 6.25%, if you add up all the other ways the less than 1MB chain can win, it adds up to c18%.
[doublepost=1465178156][/doublepost]
Yes, if you are two blocks behind, the chances of winning within 3 blocks is 1.56%, however their is no requirement to win in three blocks. You could win in 5 blocks for example. Please note its impossible to win in 4 blocks, or any even number of blocks.
Richy_T
Well-Known Member
- Dec 27, 2015
- 1,085
- 2,741
Thanks. I'm sure I can find that
I think I just glossed over that section the first time.What do you mean by winning though? This isn't a one-time thing. The smaller hash-rate chain has to keep on winning (against the odds) while the larger hash-rate chain just keeps rolling along until it pulls ahead by a lead which is realistically impossible to defeat. This is why your categorization of an attack is problematic. You assume attack parameters which applies a short-term highly concentrated conflict rather than a long-term community-led permanent change (It's a marathon, not a sprint.) and that it would produce an adverse effect (transactions would be identical between chains).
Last edited: