Will a shorter block time be helpful or harmful to Bitcoin and its future?

[jaywaka2713]

If bitcoin is attacked would a 2 minute block time make mining via Tor or something similar difficult? I don't know why anyone would mine via Tor, but yes. Tor mining would be the largest originator of Orphan blocks. Another minor point is that with block rewards 1/5 the size they will be by the current progression the rounding errors would start sooner and even less total coins would eventually be produced True, rounding errors would only occur after more than 4 block halvings, which would take about 16 years. (assuming we don't add more than 8 places after the decimal point). We will continually move down decimal points as Bitcoin's value increases. When the satoshi is worth a dollar, it will have happened.

[amincd]

If bitcoin is attacked would a 2 minute block time make mining via Tor or something similar difficult?

From what I understand about Tor, yes it would make it more difficult. Shorter block times would cause more work to be lost to latency, which would affect those with slower connections more than others. Someone more familiar with Tor, and in particular Tor mining, should answer this though.

Another minor point is that with block rewards 1/5 the size they will be by the current progression the rounding errors would start sooner and even less total coins would eventually be produced (assuming we don't add more than 8 places after the decimal point).

Good point. I think increasing the resolution of bitcoin amounts would be widely supported protocol change.

[jaywaka2713]

Another minor point is that with block rewards 1/5 the size they will be by the current progression the rounding errors would start sooner and even less total coins would eventually be produced (assuming we don't add more than 8 places after the decimal point).

Good point. I think increasing the resolution of bitcoin amounts would be widely supported protocol change.

I don't know too many people in support of raising the 21 million possible Bitcoins.

[amincd]

That wouldn't increase the amount of bitcoins. It would only allow more precise decimal fractions of bitcoin to be transferred.

e.g. instead of being limited to transferring 5.23 uBTC, you would be able to transfer 5.2345678 uBTC.

[jaywaka2713]

That wouldn't increase the amount of bitcoins. It would only allow more precise decimal fractions of bitcoin to be transferred.

e.g. instead of being limited to transferring 5.23 uBTC, you would be able to transfer 5.2345678 uBTC.

Ohh, I misunderstood what you said. Yes, that will eventually happen if/when the satoshi or uBTC is worth something (e.g. $1)

[amincd]

Even when more confirmations are required, a shorter block time will make the process faster. Waiting six minutes instead of one hour for six confirmations is an advantage.

That's not how confirmation security works, and it's been pointed out before in this thread.

I've explained in this thread why shorter block times do not increase the likelihood of an attempted double spend attack being successful, with the caveat that they make >50% attacks slightly easier and attacks using rented hardware less costly.

If I've missed something in my previous post:

https://bitcointalk.org/index.php?topic=232297.msg2470639#msg2470639

Please let me know.

*Edit, I guess one thing I haven't mentioned is that a shorter block-time block provides a smaller block reward and is therefore more expendable in an attack.

For example, if the block reward is 25 BTC for a 10 minute block, it would have to be reduced to 2.5 BTC for a 1 minute block. If you have 10% of the network hashrate, and were attempting to double spend transactions with 1 confirmation, you would have to give up 9 confirmations worth 225 BTC to successfully pull off an attack in the 10-minute blockchain, while you would only give up 22.5 BTC (9 X 2.5 BTC) in the 1-minute blockchain.

While this is all true, it's also true that the likelihood of an attacker successfully double spending a 10-confirmation transaction in a one-minute blockchain is much lower than an attacker successfully double spending a 1-confirmation transaction in a 10-minute blockchain, despite both putting the same amount of block reward at risk for the attacker and both taking the same amount of time.

Also, in many cases, people would be satisfied with much lower levels of security than that provided by one confirmation in a 10 minute blockchain, and could do with one or two confirmations in a 1-minute block-time chain.

[calian]

If bitcoin is attacked would a 2 minute block time make mining via Tor or something similar difficult? I don't know why anyone would mine via Tor, but yes. Tor mining would be the largest originator of Orphan blocks.

The reason miners would resort to mining from Tor is that bitcoin would be under attack. If anyone connecting to the bitcoin network with a resolvable IP got the dreaded knock at the door then all miners would be mining through Tor.

[ranlo]

That wouldn't increase the amount of bitcoins. It would only allow more precise decimal fractions of bitcoin to be transferred.

e.g. instead of being limited to transferring 5.23 uBTC, you would be able to transfer 5.2345678 uBTC.

Ohh, I misunderstood what you said. Yes, that will eventually happen if/when the satoshi or uBTC is worth something (e.g. $1)

I think it will probably be a while before we see this happen, but I do see it happening in the future as coins get harder and harder to obtain. Supply will be dwindling off while demand will be rising.

[grue]

google kthx

Discussion is a beautiful thing.

No, it's not "Discussion" when we're going over the same points over and over. It's just noise. If you can't show the courtesy of searching before posting, don't post.

[ranlo]

google kthx

Discussion is a beautiful thing.

No, it's not "Discussion" when we're going over the same points over and over. It's just noise. If you can't show the courtesy of searching before posting, don't post.

But grueeeeeeeeeeeeeeee, using search takes too much work!

[Garrett Burgwardt]

Even when more confirmations are required, a shorter block time will make the process faster. Waiting six minutes instead of one hour for six confirmations is an advantage.

That's not how confirmation security works, and it's been pointed out before in this thread.

I've explained in this thread why shorter block times do not increase the likelihood of an attempted double spend attack being successful, with the caveat that they make >50% attacks slightly easier and attacks using rented hardware less costly.

If I've missed something in my previous post:

https://bitcointalk.org/index.php?topic=232297.msg2470639#msg2470639

Please let me know.

*Edit, I guess one thing I haven't mentioned is that a shorter block-time block provides a smaller block reward and is therefore more expendable in an attack.

For example, if the block reward is 25 BTC for a 10 minute block, it would have to be reduced to 2.5 BTC for a 1 minute block. If you have 10% of the network hashrate, and were attempting to double spend transactions with 1 confirmation, you would have to give up 9 confirmations worth 225 BTC to successfully pull off an attack in the 10-minute blockchain, while you would only give up 22.5 BTC (9 X 2.5 BTC) in the 1-minute blockchain.

While this is all true, it's also true that the likelihood of an attacker successfully double spending a 10-confirmation transaction in a one-minute blockchain is much lower than an attacker successfully double spending a 1-confirmation transaction in a 10-minute blockchain, despite both putting the same amount of block reward at risk for the attacker and both taking the same amount of time.

Also, in many cases, people would be satisfied with much lower levels of security than that provided by one confirmation in a 10 minute blockchain, and could do with one or two confirmations in a 1-minute block-time chain.

So, we agree that one ten minute confirmation represents x amount of work, yes? The way we'd have faster blocks (and thus confirmations) is to cut the amount of work required to find a block (on average). This is adjusted via the target difficulty (raise difficulty when average block time <10 minutes, lower difficulty when the average block takes more than 10 minutes to find).

So, when we make blocks happen faster, for example 10x faster (1 minute target blocks), difficulty is 1/10th the 10 minute difficulty (assuming it's a linear scale, don't recall offhand whether it is or not). This means that each block represents 1/10th the work to secure your money, and thus you need 10 1 minute confirmation to represent 1 10 minute confirmation.

So when you say we reach 6 confirmations ("max security"), you're saying that 6 1 minute blocks == 6 10 minute blocks, which as shown, is false. You have only 6/10ths the security of one 10 minute block, and only 1/10th the security of current "max security" (for all that means).

I understand the concept of a stable difficulty. What Garrett Burgwardt said was that confirmation eta is based off of difficulty. He did not specify the difficulty/hashpower relationship. Based off his statement, as difficulty increases, confirmation eta increases. I certainly hope this isnt the case. Or he was incorrectly stating that the difficulty just simply makes it harder to confirm whilst keeping the 10 minute time. Thats probably so. My original suspicion about his statement is incorrect, and I now understand what he was saying.

My bad, I wasn't precise enough explaining my thoughts (and how bitcoin works), and I certainly didn't mean to prevent you from thinking bitcoin over, it's just tiresome seeing the same threads over and over again with people who don't fully grasp the fundamentals of bitcoin. Unfortunately there's no really good place to learn other than to come up with ideas and hear why they're bad/won't work/etc

[jdbtracker]

the current settings for bitcoin, the ten minute confirm times allows for the nodes to seattle any fork battles that may occur.

With a vpn and a terabit connection a major player can distribute a  double spend attack  while blocking pärt of the network with less than  50% hash power, effectivwly partitioning the network, in essence creating a perimeter of control. The 10 minutes allowes the slower network to reroute Their hashing power and overcome the speed advantage of propagating blocks.

A double spend attack has to reach a minimum of 6 confirmations,, with. 10% network hash power  that. Is  a  10%. Chance of getting a succesful fork, the next round drops to a 1% probability of keeping the fork. The attacker is now on a losing battle  to keep the reuse going; with larger % the attacker has dwindling probability of succesfully winning the next round.And all mining profits will disappear once they terminate the attack as the network begins to repair itself. That is a loss of 25 bitcoins  per block. So whatever they are stealing better be worth a lot, plus it can only be a p2p transaction with cash involved. 275 bitcoins is a lot of money.

The ten minutes guarantees that the fork is generated. A smart business would detect the double  spend  if the vpn was not creating a perimeter of control and even then the business would probably have their own vpn listening to world wide network transactions.

Sooner or later the network will repair itself  and most high ticket items require id, not to mention delivery time so you would have to play the odds for three weeks.

Difficulty would be the same regardless if its 2 minutes or an hour, it gets corrected every 2016 blocks

[amincd]

So, we agree that one ten minute confirmation represents x amount of work, yes? The way we'd have faster blocks (and thus confirmations) is to cut the amount of work required to find a block (on average). This is adjusted via the target difficulty (raise difficulty when average block time <10 minutes, lower difficulty when the average block takes more than 10 minutes to find).

So, when we make blocks happen faster, for example 10x faster (1 minute target blocks), difficulty is 1/10th the 10 minute difficulty (assuming it's a linear scale, don't recall offhand whether it is or not). This means that each block represents 1/10th the work to secure your money,

I agree with everything so far.

and thus you need 10 1 minute confirmation to represent 1 10 minute confirmation.

I disagree with this.

The likelihood of an attacker double spending a transaction with 10 one-minute confirmations is much lower than an attacker double spending a transaction with 1 ten-minute confirmation.

Security comes from the cost of executing a successful attack exceeding the payout from a successful attack. If the probability of an attack being successful is reduced, that means the cost of executing a successful attack increases, because it requires more attempts. Double spending a 10-one-minute-confirmation tx has the same payout for an attacker as double spending a 1-ten-minute-confirmation tx, but is harder (more costly) to execute.