Why 6 confirmations?

[litehacker]

All of the exchanges have this common standard that 6 confirmations need to pass before it accepts your Bitcoins.

I'd assume that if someone is trying to double spend, one confirmation would be enough, and if someone is trying to get 51% of the market, the number of confirmations don't matter.

Why isn't one confirmation enough?

[Jonnii]

Just paranoid security

[Joshster]

Better to be safe than sorry.

[sohaib38]

I agree that 6 is a little excessive...especially because that can take a couple of hours sometimes, it seems.

[deepceleron]

It is a point where the probability of an attacker being able to "cancel" the transaction by replacing blockchain blocks is remote. The original Bitcoin whitepaper has related statistics.

http://we.lovebitco.in/bitcoin-paper/#ch11

Code:

P < 0.001
q=0.10   z=5

which means:
Blocks needed before a less than 0.1% chance a 10% hashrate attacker could double-spend: 5

[gnicolau]

In the future that means it will be necessary more confirmations than 6 ?

thnx.

[Gabi]

No, why?

[Oldsport]

There are a million threads answering this question

[Xialla]

Better to be safe than sorry.

+1, love this adverb.

[DannyHamilton]

Better to be safe than sorry.

+1, love this adverb.

Which adverb?

I see:

• Implied subject/noun (It)
• Implied verb (is)
• Adjective (better)
• Infinitive (to be)
• Adjective (safe)
• Preposition (than)
• Implied infinitive (to be)
• Adjective (sorry)

Am I missing something?  Where's the adverb?

[cokacpkalqu91]

all the system was designed to be extremly secure.
lets stay like that..

there are also other coins with less security

[palmcoins]

Great info

[DeathAndTaxes]

We only require three confirms.

[egorisek36]

3 confirms in what coin ?

[sgrunger]

Only one confirmation is risky.
Cryptocurrency, being a peer-to-peer supported commodity, has to have more than 1 valid opinion in order to be accepted network-wide.
And 6 confirmations isn't that long IMO.

[DeathAndTaxes]

3 confirms in what coin ?

Bitcoin.  Meni posted a whitepaper of confirmation security.  For most transaction values 6 confirms is an overkill.  Satoshi never indicated 6 confirms has some magical significance.  Merchants should set confirmation policies based on the likelihood of an attack and the potential value lost if an attack is successful. 

[DeathAndTaxes]

Only one confirmation is risky.
Cryptocurrency, being a peer-to-peer supported commodity, has to have more than 1 valid opinion in order to be accepted network-wide.
And 6 confirmations isn't that long IMO.

Once again there is no one size fits all.  If you are selling $20 steam games 1 confirm is likely plenty.  An attacker is taking a significant risk attempting to double spend, the risk is in potentially orphaned blocks.  Blocks have economic value and someone with TH/s of computing power has real costs.  The benefit of a double spend ($20 game stolen) is outweighed by the cost of failing (orphaned blocks worth $3,000 each).

[TomUnderSea]

... (orphaned blocks worth $3,000 each).

Stupid question.  How did you assign the $3,000 value to an orphaned block?

[DeathAndTaxes]

... (orphaned blocks worth $3,000 each).

Stupid question.  How did you assign the $3,000 value to an orphaned block?

Its not a stupid question.  A miner who can produce a double spend block can also potentially just produce a "legit" block.  25 BTC * $120 = ~$3,000 per block.  There is a real COST for a miner to attempt (with <51% of hashpower) to extend an alternate chain.  The risk is that the rest of the network builds a longer chain and not only does the miner not successfully double spend any blocks produced in the attempt become worthless as orphans.

That cost is what gives the network security.  That cost also means that waiting 6 confirms before transfering a $20 game about as silly as keeping a 1 oz gold bullion in a $50,000 safe guarded 24/7 by fully armed security detail. 

[doobadoo]

... (orphaned blocks worth $3,000 each).

Stupid question.  How did you assign the $3,000 value to an orphaned block?

Solved block = 25 BTC reward * $120/BTC ~ $3,000

An attacker must go back and rework the block that just posted and remove the transaction he's trying to double spend. he must get lucky beating the rest of the network which is ahead and working on a block solution on top of the old block.

Chances are attacker never catches up and any blocks he did find become orphaned (never adopted b/c he's just too slow).  That compute power could have been used to score a real block by building on the valid block.  Cost is 25 btc per fake block you waste time trying to build.  Satoshi figured if you have 10% of the network hash power (which is A LOT) there is a less that 0.1% you could do this if you tried going back 5 blocks.  That is use 10% of the has power to score 6 blocks before the other 90% could score just 1.