Sure, you can speed up confirmations, but you do so at the cost of decentralization and security.

Decreased block time and/or bigger blocks requires more bandwidth and processing power. You then make it less profitable for smaller miners to continue mining, and centralize the hash rate to larger miners.

You also cause increased forks and stale blocks. With the block time being so short, then the chance of two miners mining a block at the same height increases dramatically, as does the chance of both these chains being extended. You therefore increase the risk of double spends, as well as significantly increase the amount of wasted hashrate in the mining process.

And because of all of that, your security is much worse. 1 new confirmation is nowhere near as secure as 1 old confirmation. Double spends are common. Much of the hashrate is wasted, making 51% attacks by the few remaining large miners much easier.

BTC  recommend 3 confirmations.  up to 30 minutes at a 10 minute blockspeed
LTC   recommend 6 confirmations   up to 15 minutes at a 2½ minute blockspeed
Doge recommend 6 confirmations   up to   6 minutes at a 1 minute blockspeed

Any increase in the above by an exchange is a random increase by an exchange
and unnecessary except as an artificial PR trick on the unknowing.

Satoshi kept the average block interval at around 10 minutes and he may have his reasons to do so. There are altcoins such as Doge, which have a lower block interval. I strongly believe that the relatively large gap between the blocks may be one of the ways to enhance security of the Blockchain. If someone steals a large number of coins from an exchange or any other business, then the victims have enough time to inform everyone, so that the BTC to fiat conversion can be blocked.

Mining is already fairly centralized and smaller miners tend to just use a pool for it. It isn't as feasible to be running your own server at home, hosting it in a data center with greater bandwidth and latency is better. Bandwidth requirement does not increase for the mining farm itself then.
Given the fact that connections has increased so much and that the bandwidth has improved as well, doesn't the argument weakens? Miners almost always attempts to connect to each other directly to avoid stale blocks. The lesser hops there are, the lesser chance of stale blocks. The situation of having more stales is less prominent than when Bitcoin didn't have these many optimizations.

Probability increases assuming the blocks has to propagate through several hops though, which is how most forks are formed nowadays, with at least one of the pool being less known. But wouldn't it make sense for some compromise as well?
Depends on the probability of fork occurance. Wouldn't it be more secure primarily because it takes more work to get a confirmation with a 10 minute rather than a 1 minute block interval?

But we still have stale blocks with a frequency of somewhere around 0.1% with a block time of 10 minutes. Decreasing that to 1 minute would increase the number of stale blocks exponentially.

The probability of finding a block in time x is given by the equation 1 - e^(-x/y), where y is the expected block time. If we take x = 10 and y = 600, then the chances of finding a block with 10 seconds currently is around 1.65%. If you drop the block time to a minute and make y = 60, then that chance increases to 15.35%. That's a huge increase in stale blocks and a huge increase in wasted hash power.

Elon doesn't actually care about the energy consumption (not that it is a major problem to begin with). It was just an easy target for him to use to manipulate the price. The dude launched a car in to space for crying out loud.

Currently watching. These discussions are probably my favourite ones.

Did you just make up these?

The incident won't be happening every day. And who told you that 6 confirmations are considered secure for Litecoin? Assuming that 1 Bitcoin confirmation is 4 Litecoin confirmations then it's 6*4 = 24. Dogecoin would be 6*10 = 60, because 1 Bitcoin confirmation is equivalent to 10 Dogecoin confirmations in terms of time.

If it ever happens once, it'll be a strong evidence of what is secure and what's not. I consider that response really casual, no offense. I'm personally fine with one confirmation, but if I ever wanted to transact huge amounts, I wouldn't risk it. Recently, the chain was reorganized because two miners mined the same block at the same time. The transactions on the abandoned block are now invalid.

Something must go wrong. This site calculates the equivalent time of 6 Bitcoin confirmations compared with other cryptocurrencies, based on their offered work. Correct me if I'm wrong, but the offered work refrains from security. Isn't the distribution of the hash rate that matters?

Back in 2017, the 50th percentile as mentioned by Bitcoinstats.com is ~2s but the 90th percentile will definitely take some time. It is to the best interest of the miners to be included as upper bound of 50th percentile. It is to the best interest of the miners to be as well connected to each other as possible, while centralization concerns are valid, it becomes less if the time it takes to propagate through the network is fairly quick and it isn't illogical to assume every miners will try to get the best possible connection.

Certain miners are more well connected than the rest as of now, either through zombie nodes or a similar implementation to FIBRE.

I don't disagree but that is assuming stales are still occurring on a regular basis. Since 10 minutes essentially represents 10 minute worth of PoW mining and thus a higher cost, doesn't it make sense for merchants to be still allowing 1 confirmations? The need for a 10 minute confirmation as a measure of security can be quite overblown at times; I find it sufficient for half or 10 times less of the current opportunity cost for something that costs $1.

Transactions are not invalid after getting mined in a stale block. In fact, large proportion of the transactions are probably included in both the blocks and thus no matter which fork prevails, your transaction has already been included in the chain. Stale blocks are not a very serious security risk, if the transactions pay enough fees,  then there is a high probability of them being included in both the blocks. It does make it easier for someone to double spend if merchants starting accepting 1-2 confirmations if stale are very prone, which makes it a risk and people can probably exploit it without much effort.


No. The cost of re-organizing the blocks is what matters. It doesn't exactly matter if I have 60% of the hashrate, I can easily do a 51% attack but there is no incentives as it is far more expensive than to just mine as an honest entity.

Wouldn't it be total_seconds * hashes_per_sec? In this case, wouldn't it be 360 * 100 = 36,000 hashes ever 6 minutes of work on average?

Their offered work isn't the same on these confirmations, I get it. But if the hash rate isn't “fairly” distributed then someone could offer more work in a certain time period and hence, more proof.

Does re-organization differs from replacing previous blocks? If you had 30% of the hash rate, it'd be easier for you to re-organize the blocks in contrast with someone controlling the 10%.

Sure, but even with a 1 minute block time you can't reliably achieve this. Waiting 1 minute for 1 confirmation is still too long for Starbucks or McDonald's, especially when that 1 minute could easily turn in to 5 minutes or more.

Yeah thanks. I used minutes instead of seconds, but the principle still stands. It doesn't really matter how fast or slow your confirmations are; what matters is how much work an attacker would have to do to reverse your transaction. If you have 6 confirmations in an hour or 60 in an hour, they still need to do an hour's worth of hashrate to reverse a transaction.