Difference Between 51% and 52% Attack

[Fizzgig]

Would some enlightened individuals please explain how having more hashing power than 51% will affect an attack?

Will the speed with which they can rewrite blocks increase?
Will having 50.001% effectively be the same as having 60%?
Could a 45% attack work if the attacker gets lucky by finding the correct hash before the other 55%?

[cypherdoc]

Would some enlightened individuals please explain how having more hashing power than 51% will affect an attack?

Will the speed with which they can rewrite blocks increase?
Will having 50.001% effectively be the same as having 60%?
Could a 45% attack work if the attacker gets lucky by finding the correct hash before the other 55%?

don't forget that nothing can be rewritten with this type of attack prior to block 140,700 as that is the last checkpoint.  this will be moved up soon.

[Etlase2]

Would some enlightened individuals please explain how having more hashing power than 51% will affect an attack?

Will the speed with which they can rewrite blocks increase? absolutely
Will having 50.001% effectively be the same as having 60%? not at all
Could a 45% attack work if the attacker gets lucky by finding the correct hash before the other 55%? yes, but long term effects are very limited

the more power you have, the more blocks in a row you can get. if you want to continually degrade the network, at least >50% is required.

many people focus on the double spending aspect of re-writing history, but the reality is that there will probably never be a big enough incentive to spend the money to make double spends. anyone who is going to spend the money to get >50% is going to attempt to take down the network, and all that is ever required is just greater than 50%. they can always re-write history, eventually, at that point. locking in blocks with each software release is just a very small band-aid and does not really do anything effective since trying to re-write "ancient" history is pointless when they can just make sure no new transactions ever confirm, or any that do will eventually be reversed.

[niko]

many people focus on the double spending aspect of re-writing history, but the reality is that there will probably never be a big enough incentive to spend the money to make double spends. anyone who is going to spend the money to get >50% is going to attempt to take down the network, and all that is ever required is just greater than 50%. they can always re-write history, eventually, at that point. locking in blocks with each software release is just a very small band-aid and does not really do anything effective since trying to re-write "ancient" history is pointless when they can just make sure no new transactions ever confirm, or any that do will eventually be reversed.

Why would anyone spend time and money interfering with confirmations?

[Etlase2]

um I did say why in the post. to take down the network. if bitcoin starts becoming a real threat to the established financial institution, it would be pocket change to them and by far in their best interest

[cypherdoc]

um I did say why in the post. to take down the network. if bitcoin starts becoming a real threat to the established financial institution, it would be pocket change to them and by far in their best interest

pocket change?  how so?  the network is multiples in terms of hashing power of the largest supercomputer on earth.  how much does one of those supercomputers cost?

[Meni Rosenfeld]

Would some enlightened individuals please explain how having more hashing power than 51% will affect an attack?

Will the speed with which they can rewrite blocks increase?

Yes, higher hashrate means the attacker can close the gap faster. With 51% or 50.001% he is guaranteed to outpace the network from any disadvantage (say 6 blocks), eventually. The guarantee doesn't tell you how long it will take, and in fact with 50.001% it would take on average about 6 years (it's random so it could be much less). With 51% it's 2 days. (These estimates are with a very simplified model, actual numbers could be different).

Could a 45% attack work if the attacker gets lucky by finding the correct hash before the other 55%?

Yes, he has a chance of success but no guarantee.

[Etlase2]

pocket change?  how so?  the network is multiples in terms of hashing power of the largest supercomputer on earth.  how much does one of those supercomputers cost?

you have 3,500 posts on this board and run some financial site. your "feign stupid" shtick is getting old.

[Meni Rosenfeld]

Would some enlightened individuals please explain how having more hashing power than 51% will affect an attack?

Will the speed with which they can rewrite blocks increase?

Yes, higher hashrate means the attacker can close the gap faster. With 51% or 50.001% he is guaranteed to outpace the network from any disadvantage (say 6 blocks), eventually. The guarantee doesn't tell you how long it will take, and in fact with 50.001% it would take on average about 6 years (it's random so it could be much less). With 51% it's 2 days.

Could a 45% attack work if the attacker gets lucky by finding the correct hash before the other 55%?

Yes, he has a chance of success but no guarantee.

So how long would it take with 55%? Also, could concurrently spaming the block chain with transactions lower the time or % necessary to overtake the network during attack?

A simplified model gives a result of 5 hours, but finding a more accurate estimate requires some work.

[cypherdoc]

pocket change?  how so?  the network is multiples in terms of hashing power of the largest supercomputer on earth.  how much does one of those supercomputers cost?

you have 3,500 posts on this board and run some financial site. your "feign stupid" shtick is getting old.

hey, i'm just asking you to define pocket change.

[Etlase2]

your title really should be changed from Hero Member to Bitcoin Information Minister.

[99Percent]

I estimate it would take about $30 million USD to reliably take down the bitcoin right now.

[cypherdoc]

I estimate it would take about $30 million USD to reliably take down the bitcoin right now.

how do u arrive at that number?

[99Percent]

I estimate it would take about $30 million USD to reliably take down the bitcoin right now.

how do u arrive at that number?

About $1500 per GHash multiplied by 15000 max current network hash rate (with some overhead) =$22.5 million
leaving $7.5 million for administrative and electricity costs.

This is just very quick estimate, obvious corrections welcomed.

[Etlase2]

your title really should be changed from Hero Member to Bitcoin Information Minister.

Being a prick doesn’t undo the fact that you made a ridiculous statement. By no definition and to no entity does millions of dollars represent “pocket change.”

really, we're going to take a gross exaggeration so out of context as to call it "ridiculous"? even though it has surely been used in the same manner many times before. am I writing for an esteemed publication or on a fucking message board?

besides, comparing the bitcoin network in integer operations when supercomputers are measured in floating point operations is well-known around here to be a misguided comparison and only goes to show that cypherdoc is, once again, trying to be deceptive to people asking legitimate questions in what I can only see as being a shill for his own personal gain. he may not like my opinion or my exaggeration, but that is not course for conflation with obvious intent to confuse and derail

[cypherdoc]

I estimate it would take about $30 million USD to reliably take down the bitcoin right now.

how do u arrive at that number?

About $1500 per GHash multiplied by 15000 max current network hash rate (with some overhead) =$22.5 million
leaving $7.5 million for administrative and electricity costs.

This is just very quick estimate, obvious corrections welcomed.

i disagree with this logic.  if the US gov't or some financial institution were to enter the market for gpu's the cost of those gpu's would skyrocket pushing the cost into the billions i'd estimate.

look at the Butterfly Lab Rig unit.  b/c of the expressed interest the price has gone from $24,890 to $29,890 and they haven't even produced one!

[cypherdoc]

your title really should be changed from Hero Member to Bitcoin Information Minister.

Being a prick doesn’t undo the fact that you made a ridiculous statement. By no definition and to no entity does millions of dollars represent “pocket change.”

really, we're going to take a gross exaggeration so out of context as to call it "ridiculous"? even though it has surely been used in the same manner in many times before. am I writing for an esteemed publication or on a fucking message board?

besides, comparing the bitcoin network in integer operations when supercomputers are measured in floating point operations is well-known around here to be a misguided comparison and only goes to show that cypherdoc is, once again, trying to be deceptive to people asking legitimate questions in what I can only see as being a shill for his own personal gain. he may not like my opinion or my exaggeration, but that is not course for conflation with obvious intent to confuse and derail

trying to be deceptive?  i just asked you to explain your claim of "pocket change".  you are one sensitive dude.

shill for personal gain?  i am just one of many who have a vested interest in Bitcoin around here.

how about you and Encoin?  just who has a vested interest?  how's that going btw?

confuse and derail?  you have problems.

[cypherdoc]

your title really should be changed from Hero Member to Bitcoin Information Minister.

Being a prick doesn’t undo the fact that you made a ridiculous statement. By no definition and to no entity does millions of dollars represent “pocket change.”

really, we're going to take a gross exaggeration so out of context as to call it "ridiculous"? even though it has surely been used in the same manner many times before. am I writing for an esteemed publication or on a fucking message board?

besides, comparing the bitcoin network in integer operations when supercomputers are measured in floating point operations is well-known around here to be a misguided comparison and only goes to show that cypherdoc is, once again, trying to be deceptive to people asking legitimate questions in what I can only see as being a shill for his own personal gain. he may not like my opinion or my exaggeration, but that is not course for conflation with obvious intent to confuse and derail

It sounds to me like you have a previous conflict with cypherdoc. I don’t have a retort for that situation.

no we don't.  Etlase has conflicts with many ppl around here.  i'm just the latest.

[CleverMiner]

~~~

Cypherdoc, get a clue or gtfo. https://bitcointalk.org/index.php?topic=52166.msg623111#msg623111

[Minor]

I estimate it would take about $30 million USD to reliably take down the bitcoin right now.

how do u arrive at that number?

About $1500 per GHash multiplied by 15000 max current network hash rate (with some overhead) =$22.5 million
leaving $7.5 million for administrative and electricity costs.

This is just very quick estimate, obvious corrections welcomed.

i disagree with this logic.  if the US gov't or some financial institution were to enter the market for gpu's the cost of those gpu's would skyrocket pushing the cost into the billions i'd estimate.

look at the Butterfly Lab Rig unit.  b/c of the expressed interest the price has gone from $24,890 to $29,890 and they haven't even produced one!

If an entity decided to spend $30M to attack or take control of Bitcoin, they would NOT got out to buy GPUs in retail stores.
They could design a 65nm ASIC for a couple of millions $ of NRE and then produce a ton of them probably for less than 100$ per GH/s (wild guess based on 45nm FPGA based systems going for close to $1 per MH/s  in single quantities today. A dedicated, mass produced ASIC design should be able to be 10 times more cost efficient than a retail FPGA design).
Heck, if anyone wants to give me $5M, I'm willing to give it a shot.