Topic: Will a shorter block time be helpful or harmful to Bitcoin and its future? - page 3. (Read 4414 times)

I've been thinking about this lately - the implications of block production time on security. It's discussed a bit in the following paper:

Rosenfeld, M. (2012). Analysis of hashrate-based double-spending, 1–14. Retrieved from https://bitcoil.co.il/Doublespend.pdf


Regarding orphans - someone mentioned that the time it takes a message to get from one end of the network to the other is 10 - 20seconds (don't know where that number comes from) but the indicator of how many orphans there are going to be is some function of the time it takes for a new block to propagate and the proportion that is of the block production rate.

e.g. 30 second block time, 30 second propagation time results in huge orphan rate.
10 minute block time, 2 second propogation time results in tiny orphan rate.

The issue of orphan transactions as well is slightly negated by quicker block production time.

e.g. it doesn't matter as much if your transaction gets orphaned once in a 30 second block rate, because it's only delayed by 30 seconds, but in a 10 minute block rate, that's 10 minute delay at least.

I think it should be left alone 10Mins was what the creator had intended lets not start manipulating the code thats what alt coins are for, let them have this advantage at least e.g Litecoin with 2.5mins 

don't you need more confirmations to get the same security with shorter block times, also, I'm not sure how this would go for mass uptake

eg, when I shop, it takes the visa card about 1-3 seconds to confirm (its a pre-auth I think), and I am not sure a 1 sec is really possible with current tech, eg you would need 0.1 sec confirm to get 10 confirms and you still do have very poor security,

I may be wrong, but it a network propagation issue, that I think there is a solution by some sort....

Harmful in the future

Conf time is indirectly determined by difficulty, with difficulty roughly determining how many hashes are required to solve a block (thus, roughly how much time is required to solve a block).

Difficulty is determined by the target confirmation time (10 minutes) and the time it's taken to solve the past 2016 blocks. With 2016 blocks each with a target confirmation time of 10m, Difficulty should adjust about every two weeks, and is meant to adjust so it always takes roughly 10 minutes to solve a block. Because we're in a period of rapid network hashrate growth with the release of ASICs, it should actually take significantly less than 10 minutes to solve a block the closer we get to the next time difficulty corrects, but also means difficulty will correct sooner than the target of two weeks. (I'm sure someone has stats, but for the past month, it's probably been something like 10-12 days for each set of 2016 blocks instead of the target 14 days)

A target confirmation time of 5m instead of 10m would require halving the difficulty.

Trivia: Cryptocoins are subject to a kind of "spiral of death" if hashrate plummets. If, say, 75% of hash power dropped from the network a day after the difficulty is adjusted, it would take (13*3), or 39 days for difficulty to adjust to that sudden drop. Confirmations would take an excessively long time to confirm, possibly eroding faith in Bitcoin and causing a kind of "hash run," where even more hash power drops off until it's nearly impossible to confirm transactions anymore, leading to abandonment (or maybe some kind of fork).

Trivia2: Changing the target confirmation time is a protocol change and relatively difficult to have accepted.

Trivia3: I'm tired as dog-shit, and someone else should be much more able to assist you than I (and hopefully correct any misinfo I gave). Good night!

No it will only increase Bitcoins value and be beneficial.

OK so if I understand you right, if we drop block time by half to 5 minutes, and doubled difficulty, would that keep the hashrate necessary stable? I'm not even sure that is correct, as I didn't think block time effected hashrate requirement to crack a block. Block time is just a variable correct? Or does block time tie directly into difficulty?

Sorry. I write like a shithead, sometimes.  I'll try to clarify some points I wrote particularly poorly.

When I was talking about expense to confirm double-spends, I was referring to the cost to compromise the Bitcoin network by making up a "dangerous" percentage of the total network hash rate. Network hashrate is dramatically higher now than when 6-conf suggestion was made (that is, when it was decided that 6 confirmations would be implied as "safe" by the Bitcoin client's graphics). It would likely cost millions, now, to get 3 confirmations on a double-spend, whereas it may've taken just a few thousand USD in hardware or hashpower leases far in the past. In this sense, the network is dramatically more secure, now, as it would be extremely expensive to successfully pass off a double-spend as legitimate through multiple block-solvings.

By conf time, I was just referring to the target confirmation time of a block (currently 10m). If you reduce it (say, to 5m), you are decreasing the amount of hash-power required to solve a block (exactly by half compared to what's required with a 10m target conf time). By doing so, it would be twice as easy to pull off a one-conf attack, which many would consider legitimate for smaller transactions. (but, again, this is still now prohibitively expensive even if it's twice as easy to pull off than now... one-conf probably won't be accepted as "legitimate" by most merchants selling something of significant value, so it's kind of a moot point).

The orphan argument I only picked up from others -- I haven't seen stats on it, so I'd take your word on it since I don't have any counter-examples.

ETA: Btw, don't consider me some type of authority on this stuff. I've always shied away from the technical underpinnings of BTC -- ADD, maybe. Bigger block sizes (more transactions allowed per block) might be considered a favorable option of ensuring speedy transaction processing because it doesn't introduce risks at all. Increasing block size 2x would have the same (eventual) effect as having 2x as many blocks, but wouldn't introduce the arguable risk of making merchants more vulnerable to one-conf double-spends (keeping in mind that one-conf double spends are 2x as computationally easy, and thus effectively 2x as cheap with half the conf time).

Actually, after the extensive testing of various altcoins, 2 minutes for a block time will not create many orphans at all, if any. After reading your post, it is obvious to myself that I do not fully understand what a confirmation is. Can you send me a link for me to read concerning confirmations?

You also said something about dropping the confirmation time. I was not aware such a variable existed. Would that decrease the integrity of the confirmations?

Faster block conf times leads to a higher rate of orphaned blocks, increases the likelihood of dishonest folks from succeeding in double-spends in certain circumstances, and creates a larger blockchain.

A confirmation is not a guarantee of authenticity. Decreasing conf time proportionally decreases the "value" of the confirmation. Decreasing confirmation times would be accounted for in services requiring confirmations, where required confirmation times would be increased to completely negate the decreased conf times. (for example, Gox requires 6 confs -- reducing conf time to 5m would result in Gox requiring 12 confs)

... That said... to compromise the Bitcoin network back when the six-conf suggestion was suggested, it was dramatically less expensive to have double-spends confirm. Three confirmations is now more than adequate for all transactions under hundreds of thousands in USD. Going beyond is really just (inadequately) protecting yourself against a fork. Maybe it is time to discuss lowering the target conf time?

ETA: Devs may also be trying to encourage Bitcoin transactions to occur internally on central servers to keep "casual" transactions off the blockchain... OTOH, they seem to be in general consensus that the block size should be increased in the near future. I'm sure one'll come in soon and tell us what's up.

I too am very interested in their opinions. Block time is one of Bitcoins big issues for massive adoption, but is one of the easier issues to fix as well. The repercussions of such a change have to be discussed though.

EDIT: Fixed Poll. Yes/No were invalid answers to the question LOL.

I guess shorter time should be better. But I want to listen the opinion of the old members.

Almost everyone has seen the alternate currencies section of this forum. There has been just about one hundred coins developed, all with a variable range of block times. We have seen one benefit from this mass testing, and that benefit is faster block confirmation times.

I, a two year long Bitcoin user, have experienced the issues with the current 10 minute block time. There has been endless discussion about making zero-confirmation transactions safe, but what if a lighter solution is better? Why not reduce the block time? Now, I understand that a block time of 30 seconds does nothing but damage, and will not propose that, but what about a block time of 2 minutes? That will speed up transactions by 500%, and will make the hour-long transaction something of the past.

That 2 minute block time is long enough to prevent any orphans at all, and the network has been very well developed up to this point. However, changing the block time changes other things as well. One such example would be block reward. At the time of writing, the block reward is set to 25 BTC. So if the block time were to be changed to 2 minutes, the block reward would need to be changed to 5 BTC.

Yes, there are drawbacks to this change. One such drawback would have a direct effect on those who solo mine. Mining is really a game of chance. Cracking a block would only give you one fifth of the reward. There are also pros as well. This would allow pools to have higher odds in cracking blocks as well, as blocks are created 5 times faster. Block reward would continue to halve along with the current plan.

My main point is that transactions are simply too long. The one advantage the banking system has over us is instant transactions. Reducing the time required for about 6 confirmations from half an hour/hour to around 10 minutes would be very helpful. Also makes 1 confirmation transactions very fast as well. Thoughts?