Is the BTC network broken?

[MRKLYE]

These block times range from constipation to diarrhea... My understanding was 10 minute blocks?
What has happened?

[Meni Rosenfeld]

10 minute average (slightly less when hashrate is increasing), actual time between blocks varies and follows the exponential distribution.

[DeathAndTaxes]

These block times range from constipation to diarrhea... My understanding was 10 minute blocks?
What has happened?

Your understanding was wrong.  Bitcoin mining is random and follows a poisson distribution.  The AVERAGE time between blocks will be 10 minutes when the network is neither growing or shrinking.  The time between any two blocks is random.

[MRKLYE]

These block times range from constipation to diarrhea... My understanding was 10 minute blocks?
What has happened?

Your understanding was wrong.  Bitcoin mining is random and follows a poisson distribution.  The AVERAGE time between blocks will be 10 minutes when the network is neither growing or shrinking.  The time between any two blocks is random.

Thanks for the explanation!

[Meni Rosenfeld]

Bitcoin mining is random and follows a poisson distribution.

Bitcoin mining follows the Poisson process. Number of blocks in an interval follows the Poisson distribution.

[MysteryMiner]

These block times range from constipation to diarrhea... My understanding was 10 minute blocks?
What has happened?

When dealing with bitcoin constipation try taking some block erupters!

Seriously the average time is 10 minutes, the time between blocks can be between few seconds and as much as 9 hours. To get at least 1 confirm is less than any other payment alternative. If 10 min average is not enough and people will get guaranteed in 10 seconds somebody will come and complain about need to wait 10 sec.

[RodeoX]

Just out of curiosity, does anyone happen to know the shortest and longest periods between blocks thus far?

[MysteryMiner]

Just out of curiosity, does anyone happen to know the shortest and longest periods between blocks thus far?

There was a thread long ago about longest time between blocks. If I remember it was like 9 hours.

[RodeoX]

Just out of curiosity, does anyone happen to know the shortest and longest periods between blocks thus far?

There was a thread long ago about longest time between blocks. If I remember it was like 9 hours.

Wow, I'm guessing that made a few people nervous.
Thanks MM.

[Meni Rosenfeld]

Just out of curiosity, does anyone happen to know the shortest and longest periods between blocks thus far?

Better yet, where can we found tabular data of blocks and their vital data (e.g., timestamp and difficulty) from which this can easily be deduced? The closest to this I found was blockexplorer nethash, but the minimum allowed is steps of 5 blocks.

"Shortest interval" would be somewhat undefined since published timestamps don't have to be synchronized, you can have negative intervals.

As for longest, if 254508 blocks are distributed exponentially with mean 10 minutes, you'd expect the max to be 130 minutes.

Just out of curiosity, does anyone happen to know the shortest and longest periods between blocks thus far?

There was a thread long ago about longest time between blocks. If I remember it was like 9 hours.

That's virtually impossible unless there was something very wrong with the network. If it's distributed as usual the chance of that is one in a trillion trillions.

[flatfly]

block 15324 (May 2009), 25 hours:
 http://blockchain.info/en/block/00000000720e7a19e31ca6cc9d2364705b49590f0ede8b0356cb46a8e486ab8f

and a bunch of others...

[Meni Rosenfeld]

block 15324 (May 2009), 25 hours:
 http://blockchain.info/en/block/00000000720e7a19e31ca6cc9d2364705b49590f0ede8b0356cb46a8e486ab8f

and a bunch of others...

Yeah, probably Satoshi was the only miner and he turned off his computer... Indeed, something very wrong with the network.

[DeathAndTaxes]

block 15324 (May 2009), 25 hours:
 http://blockchain.info/en/block/00000000720e7a19e31ca6cc9d2364705b49590f0ede8b0356cb46a8e486ab8f

and a bunch of others...

Yeah, probably Satoshi was the only miner and he turned off his computer... Indeed, something very wrong with the network.

Due to the low number of active nodes and the fact that hashing power didn't even meet difficulty (the expected time between blocks was ~30 minutes) blocks from the first year aren't really useful. Anecdotal I remember waiting 2 hours for the 6th confirm (not 6 confirms just one more confirm to make 6) so MtGox would credit my account.

It would be useful if someone pulled the timestamps from the blockchain and put them in table form.  The only problem is that since the network allows such loose timestamps any 1 block distribution is going to have "weird" numbers (like negative block times).  However the results should be cleaner if one looked at multiple bloks.  i.e. the distribution of a 3 block or 6 block interval.  That might be more useful in a lot of circumstances (like seeing how rare my deposit at MtGox frustration was).

[Meni Rosenfeld]

block 15324 (May 2009), 25 hours:
 http://blockchain.info/en/block/00000000720e7a19e31ca6cc9d2364705b49590f0ede8b0356cb46a8e486ab8f

and a bunch of others...

Yeah, probably Satoshi was the only miner and he turned off his computer... Indeed, something very wrong with the network.

Due to the low number of active nodes and the fact that hashing power didn't even meet difficulty (the expected time between blocks was ~30 minutes) blocks from the first year aren't really useful. Anecdotal I remember waiting 2 hours for the 6th confirm (not 6 confirms just one more confirm to make 6) so MtGox would credit my account.

It would be useful if someone pulled the timestamps from the blockchain and put them in table form.  The only problem is that since the network allows such loose timestamps any 1 block distribution is going to have "weird" numbers (like negative block times).  However the results should be cleaner if one looked at multiple bloks.  i.e. the distribution of a 3 block or 6 block interval.  That might be more useful in a lot of circumstances (like seeing how rare my deposit at MtGox frustration was).

Well, for 6 block intervals you can use http://blockexplorer.com/q/nethash/6. 2nd field is Unix timestamp.

For reference, the time until 6 blocks are founds follows the Erlang distribution, with shape parameter 6.

[twmz]

Better yet, where can we found tabular data of blocks and their vital data (e.g., timestamp and difficulty) from which this can easily be deduced? The closest to this I found was blockexplorer nethash, but the minimum allowed is steps of 5 blocks.

Data on all blocks can be obtained by the blockchain.info API, but you'd have to make multiple API calls to get all the data (they have no API that dumps everything in one call).  Probably the blocks-in-one-day API would be the easiest to access.  Just have to make 1 call for each day since the network went live (which is many fewer requests than 1 per block-height).

[MysteryMiner]

I have blockchain with client version from 0.20.2 incrementally upgraded to 0.4.1 on harddrive since February 2011 till November 2012. Is it possible to get additional data from it including when each block got received? I guess there might be additional data in BerkleyDB files but I'm not sure.