Confirmation

[Walter Rothbard]

Confirmation means reducing the likelihood that a transaction can be double spent down to an acceptably unlikely level.

There is a lot of math in the Satoshi whitepaper about how the likelihood that a transaction can be double spent reduces with each block the transaction is buried under.  All of this math is based on average 10 minute blocks and the parameters with which Bitcoin was implemented.

If you change all those parameters, like the timing of the blocks, you will also be changing the difficulty of rewriting the block chain and thus changing the likelihood that a transaction can be double spent.

You cannot merely make blocks faster and claim that your blockchain has "faster confirmations."  You have to do the math!

[jasemoney]

have you done the math? I haven't but once a transaction is included in the chain it'd have to be 51%'d to be overwritten as a double spend from a chain height before you did the first spend. Any coin that isn't constantly forking itself should be pretty safe that a confirm = legit.

[Walter Rothbard]

I haven't but once a transaction is included in the chain it'd have to be 51%'d to be overwritten as a double spend

That would be trivial to do on a lot of alt-chains, where far more hashpower is available on other chains that could be pulled over for long enough to do this.

[Walter Rothbard]

Section 11 calculates the worst case under attack. Typically, 5 or
10 blocks is enough for that. If you're selling something that
doesn't merit a network-scale attack to steal it, in practice you
could cut it closer.

http://satoshi.nakamotoinstitute.org/emails/cryptography/14/

[Walter Rothbard]

Never thought of that, is it true?

Read the Satoshi whitepaper and find out how it works!

[Willisius]

Confirmation means reducing the likelihood that a transaction can be double spent down to an acceptably unlikely level.

There is a lot of math in the Satoshi whitepaper about how the likelihood that a transaction can be double spent reduces with each block the transaction is buried under.  All of this math is based on average 10 minute blocks and the parameters with which Bitcoin was implemented.

If you change all those parameters, like the timing of the blocks, you will also be changing the difficulty of rewriting the block chain and thus changing the likelihood that a transaction can be double spent.

You cannot merely make blocks faster and claim that your blockchain has "faster confirmations."  You have to do the math!

From Meni Rosenfield's Analysis of hashrate-based double-spending (link):

The probability of success depends on the number of confirmations and not on the amount of time waited. An alternative network with a different time constant T₀ can thus obtain more security with a given amount of wait time.

Consider the math done. (I'm not being lazy, I'm just way more confident in his math than my own.)

[iGotSpots]

If only coins used chain "time" and not real life seconds

If only...

[Viper1]

Confirmation means reducing the likelihood that a transaction can be double spent down to an acceptably unlikely level.

There is a lot of math in the Satoshi whitepaper about how the likelihood that a transaction can be double spent reduces with each block the transaction is buried under.  All of this math is based on average 10 minute blocks and the parameters with which Bitcoin was implemented.

If you change all those parameters, like the timing of the blocks, you will also be changing the difficulty of rewriting the block chain and thus changing the likelihood that a transaction can be double spent.

You cannot merely make blocks faster and claim that your blockchain has "faster confirmations."  You have to do the math!

From Meni Rosenfield's Analysis of hashrate-based double-spending (link):

The probability of success depends on the number of confirmations and not on the amount of time waited. An alternative network with a different time constant T₀ can thus obtain more security with a given amount of wait time.

Consider the math done. (I'm not being lazy, I'm just way more confident in his math than my own.)

"The time constant might be relevant, if we assume that the attacker cannot sustain his hashrate for a prolonged period of time."

6 confirmations with 10 minute block times would mean you'd have to maintain enough of a hashrate for over an hour.  6 confirmations with 1 minute block times means you only have to maintain it for just over 6 minutes or so.  Course if you're actually relying on the shorter time frame to be successful, you probably don't have the hashrate to succeed anyway. lol

[Este Nuno]

If the time is a constant to achieve the same probability of a successful secure transaction, then shorter confirmations are still better obviously. The only times when you want full security is going to be times when the transactions are large enough to matter then you're waiting the same amount of time either way.

The only question in my mind is how short can we get? 1-2 minutes seems to be the general consensus, but it's not clear how much wasted hash(orphans) is an acceptable amount.

10 minute confirms seem to have been a small error on Satoshi's part. In my opinion it comes down to whether you value miners over users. And clearly users are what matters when we're talking about a currency.