Quote from: rudi on February 27, 2014, 05:11:56 PM

Quote from: gog1 on February 27, 2014, 04:58:10 PM

If we find 3 blocks in 7 days, that's avg of 3/7 blocks per day and if assume Poisson process, we have the following:
blocks found / probability
0 / 0.6514
1 / 0.2792
2 / 0.0598
3 / 0.0085
4 / 0.0001+

Going back all the way to mid Dec, I cannot find any 24 hour period where we have >3 blocks.

Almost.
We need to use the cumulative distribution function.
Then, assuming a rate of 3/7 blocks found per day, we get:

Chance of finding more than 0 blocks per day: 34.856094 %
Chance of finding more than 1 block per day: 6.937277 %
Chance of finding more than 2 blocks per day: 0.954674 %
Chance of finding more than 3 blocks per day: 0.100016 %

This is not just variance.

(In python:
import scipy.stats
1. - scipy.stats.distributions.poisson.cdf(blocks_per_day_new, blocks_per_day_old))

At the end of the day, there are still uncertainty until we see announcement and things come along. The recent could be
1) luck
2) friedcat is testing 'sample' miners
Even if things look good, there's always potential for operational issue as you do not control the manufacturing & assembly plant.