Bitcoin Forum

The oldest block said only "11 minutes" and by the time I decided to grab the data it was 12 minutes:

305068   2 minutes   21   33.05 BTC   Discus Fish   12.18164062
305067   2 minutes   22   186.91 BTC   GHash.IO   8.73144531
305066   3 minutes   146   660.13 BTC   GHash.IO   84.96386719
305065   7 minutes   67   1,599.99 BTC   GHash.IO   124.86132812
305064   8 minutes   181   2,296.68 BTC   23.253.207.62   168.10058594
305063   12 minutes   154   3,689.36 BTC   GHash.IO   324.90136719

My last transaction confirmed so fast it was amazing.  :)
I know this can happen, but isn't it rare for BTC?

I haven't heard of a six block solve in less than ten minutes
I think the average is 2 to 3 at most unless their is a big hash change
Think that means the difficulty will rise faster now or nothing will be solved for a while ?
Edit see 305069
https://blockchain.info/

This the reason why difficulty has been climbing month after month at an exponential rate. I blame it on too much hashing power on a single pool. I am happy I am not a miner, ROI seems further and further away.

It means that hashing power had to almost double, no? :O

Bitcoin on gold rush would be a better on with extra testosterone.

If the network solved 4032 blocks in the time for 2016 (~2 weeks) then it would mean it doubled.   There is a huge amount of variance.  The average time between blocks over a period of 6 or even 48 blocks is mostly random noise.  Even for 144 blocks (~1 day) there is huge variance.

It's mix of hashing power going up, and little bit of luck (dont underestimate luck). You can see how diff and hash power is raising (or not) here: https://bitcoinwisdom.com/bitcoin/difficulty

Next difficulty might be the last one for my old blades.  :(

"Random noise" is correct, but I wonder if this has ever happened before?  (other than the really early days of mining)
I know I've never had a BTC transaction completely confirmed so fast.

I recall having 4 confirmations in a couple of minutes. That was funding my btc-e wallet to do a quick trade, so the speed was really appreciated :).

I pulled out after last difficulty. The heat generated in the summer didn't justify the tiny quantities of BTC.
I did keep a Jalapeño and decided to solo mine with it. Odds are insanely stupid, but heck! Let's play!

Oh, and I kept my AMT miner - Those things are unsellable!

I also bailed out of my pool
The hashing escalation completely killed my RoI

https://i.imgur.com/Y5AGd.gif

don't look the clock ... look the graph : https://bitcoinwisdom.com/bitcoin/difficulty

I don't understand why difficulty increases would be different whether the hashing power is concentrated in a single pool or spread out among many pools. Can expand more on that?

it doesn't usually happen that fast... (and sometimes I have to wait like 20 minutes to get my transaction confirmed)  ;) They all worked for you lol

It is not. It's eather that poster before you has no idea what he's talking about, or tehere was misunderstanding in what he wanted to type and what he actualy typed. Hash speed is hash speed, in 1 pool or 10 of them. This blocks were just luck, nothing more. Hash speed is going up steady as it used to go up for months now.

How did they actually count the *confirmation ?

I can see 50-250 % noted something like that on the detail of my transaction

In the early days of mining (2009) there actually was insufficient hashrate to meet the min difficulty of 1.  The average time between blocks varied from 12 minutes to 16 minutes.  If the protocol allowed the difficulty to be lowered below 1 it would have there was so little hashrate.    So it would have been less common in the early days.

It is certainly uncommon.  I am not sure how often it has occurred in the past but here are the odds.  Mining is a Poisson distribution.  So if difficulty and hashrate were in perfect equilibrium in an 11 minute period one would expect 1.1 blocks.  The probability of getting 6 or more events when 1.1 is expected is 0.1% so it would occur on average every 1033 events or roughly once a week.  However the hashrate currently exceeds the difficulty and the time between blocks averages less than 10 minutes (currently 8:59) so this means the probability is about 0.17% and it would be expected roughly once in 606 blocks.   Now those are the odds based on pure math.  In reality higher block frequency (on the best chain) is less common because in the real world there are propogation delays and that means the shorter the interval the more instead of 2 blocks in a row you have 2 competing blocks (A -> B1 & A-> B2 instead of A -> B -> C).   One last thing is this assumes the block timestamps were actually correct it is possible the first or last one was early or late and thus the actual period of time was longer.

Still it does show why Satoshi was so conservative with the target block interval.  Imagine if Bitcoin had a 30 second block and now imagine the tx volume was so high those blocks averaged 10MB.  The network would need to propagate and sync 60MB of data to all nodes in less than 33 seconds or there would be orphans (potentially multiple).

6 blocks in 10 minutes has happened a few times before, just like 0 blocks in 60 minutes happens about once a month.  6 in 10 minutes is a bit rarer though.

I think some 40-50 days back it was really fast. Sometimes there would be 3 confirms in a couple of minutes.

...

With so many here who know more than I do,,, I hesitate to comment, but I think "DeathAndTaxes" has it right:

Probability!

Sometimes the unusual happens.

I once saw a time difference of 75 some minutes between blocks not too long ago.

***

When I get back I will have to consult my probability text on "Poisson Distribution".

This is not all that unheard of. Especially considering how fast the network is expanding.

According to the rule of infinite possibilities, there could even be 100 blocks found in 11 minutes. It's possible, just highly improbable. When difficulty adjust every 2 weeks, everything averages out according to total network hashrate.

You are correct in saying this, however the law of probabilities says that something with a certain probability should happen once every x time on average.

Having 6 blocks mined in 11 minutes happens relatively often. I am not sure exactly how often, it isn't once every two weeks, but also not once every three months.

I have had this same exact thing happen on a few occasions. Most of the time it takes significantly longer though 

Yes, variance is perfectly normal in any statistical model. People often see it as a pattern but it is simply statistical outliers.

The rapid block sequences will tend to occur right before a difficulty change if we're growing rapidly.  Look back to those increases where we had 30% jumps and the last couple of days they wold come out rapid fire.

FTFY

It happens occasionally and it is probably due to GHASH.io's majority of the hash power (notice 4 out of the 6 are theirs).

Ghash is on steroids. :D

No...it has NOTHING to do with where hash power is. On 1 pool or 1000 pools, it has absolutley  no effect. Hash power is hash power. It's all about hash power and statistical luck. As some said before, it happened before, it will happen again. Sometimes you get rapid ones, sometimes you wait for 1 hour to get 1....pure and simple luck. Nothing more, nothing less.

How far ahead of the block schedule is bitcoin overall? Is it adding up over time to put us significantly ahead of schedule? Block reward halving going to come sooner than predicted before?

If blocks were found exactly each 10 minutes, we'd be expecting a halving in January 2017, however we're more going into August 2016 (http://bitcoinclock.com/).

Since DeathandTaxes did the math for the frequency of the OPs post, I couldn't help but do the comparison of this during one of my breaks at work. Hehe.

With ideal block propagation averages:

6 blocks in 10 minutes (i.e. P[X=6] for λ=1) once every ~13.6 days.
0 blocks in 60 minutes (i.e. P[X=0] for λ=6) once every ~16.8 days.

As a note the chances of 6 or more blocks in 10 minutes (i.e. 1-P[X<6] for λ=1) is once every ~11.6 days.

So it seems that 6 blocks in 10 minutes is a bit more common than no blocks in the expected time for 6 blocks.

http://bitcoinclock.com/

This will help you alot.

...

ljudotina

Nice link!  Thank you.

More or less depending on the hash rate change until then ;)

Thanks to both you.

It may be even earlier.

I would not be surprised if it happened as early as in 2015.

It will be earlier, I was just pointing to the current estimation.

Yes, it all depends on how much hash is added to network in between diff changes. Steeper change, sooner it will get.

Just had 6 blocks within 16 minutes (ending ~17 minutes ago)

305463    17 minutes    395    2,273.73 BTC    72.83.147.125    243.36
305462    22 minutes    32    504.36 BTC    Polmine    14.85
305461    26 minutes    81    497.38 BTC    Slush    46.14
305460    27 minutes    35    469.12 BTC    GHash.IO    21.07
305459    24 minutes    432    46,996.86 BTC    BTC Guild    287.12
305458    33 minutes    128    101,239.74 BTC    Polmine    148.88

Obviously a few fast blocks is just from random odds.



So it is normal over twice (or even three times) a month?
Seeing the math is amazing thank you and DeathandTaxes for doing that.

Yes, it's normal for the six block thing to happen multiple times in a month. I know, crazy huh? That's the funny thing about probabilities and proportionality. When something has a .05% chance of happening, something in a lot of people instinctively think that is super rare. But over the scale of thousands of events such as a month of which there are around 4,300 10-minute events, that .05% event has an expectancy of occurring about 2 times.

As a side note, we would expect a 7-block event once every ~3 months, an 8-block event every ~2 years, a 9-block event every ~2 decades and a 10-block event ~2 "lifetimes" (about 187 years). Mind you though, this is assuming "normal" block propagation times (which we are running faster than at the moment).

It has to do with luck as well.

still means my coffee woulda been cold if a) coffee shop accepted bitcoins and b) waited for 6 confirms
but ya, 6 in 11 mins is not to shaby at all :)

Merchants that sell low value products like coffee could rely on 0/unconfirmed TX as it would cost a lot to attempt a double spend attack on a $5 purchase

Why the hell is it compared to anything on steroids?

Yup , new miner is really taking a lot time to cover their investment tho and it might take longer than just buying the bitcoin outright.

Most miners will never reach ROI and never will they. Miners have to pay for electric costs out of their mining proceeds, and when electric costs exceed their proceeds they will shut off their miners.