The variance of block times should arguably follow a Poisson distribution. But by casual observation I've seen blocks fast as 4 s or 1 m 15 s. The average is around 14 s since Homestead.
If it follows a Poisson distribution it would have not that huge variance I observe:

enter image description here

What is the true distribution of block times in Homestead, derived from empirical observation?


First of all it seems that I am too ignorant to understand Poissons Lamda!
At least I have some R skills left:

 Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
   1.00    5.00   10.00   14.35   19.00  155.00 

Interpretation: if you take the mean and forget the stupid outliers we count with only 10 s block time!

density block time cdf block time

The good news: 99% blocks are under 1 min

 quantile(blocktimes$blocktime, c(.9,.95,.99))
 90% 95% 99% 
 31  40  61 

The peak is around 2.9 s

enter image description here

Source code on Github


Please see these statistics from etherchain.org: https://stats.etherchain.org/dashboard/db/average-block-time?theme=light :)

  • 2
    Sorry but it shows only the daily average block time. I'm looking for the statistical distribution of block times. A man with one hand in the freezer and one in the oven can enjoy room temperature on average ;-) – Roland Kofler Jun 10 '16 at 15:41
  • added more details – Roland Kofler Jun 10 '16 at 16:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.