Quote from: amincd on June 14, 2013, 07:46:22 PM

Quote from: Garrett Burgwardt on June 14, 2013, 05:28:48 PM

Quote from: amincd on June 14, 2013, 03:40:46 PM

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).

Quote from: jaywaka2713 on June 14, 2013, 05:41:24 PM

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