I ran some quick numbers. If I'm correctly understanding the distribution, once the first block was found, the chance of finding the next 6 blocks in that time window was something like 0.0000002. However, this possibility has existed every time a block has been found. The chance of this streak happening at least once in 132000 or so blocks is greater than 99.7%. In fact, there is a very high chance that his has happened more than once. If I find the time, I could do an analysis of the whole block chain to see if this has happened before.
Is there an easy way to extract block timestamps from the chain?
If you evaluated the cdf of a gamma distribution with parameters 7 and 1/10min at 10min to get 0.0000002, and 1 - the pdf of the binomial distribution with parameters 132000 and 0.0000002 evaluated at 0 to get 99.7%, then I think it's correct. I'm too lazy to calculate it, though.
Also,
http://www.xkcd.com/882/My knowledge of statistics is a bit rusty, so I decided to look at the actual blocks. I ran a script to evaluate, from each block, how many blocks were found less than 420 seconds beforehand.
The record was
twelve.
There were 358 cases of 6 or more (ie. 7 blocks in under 7 minutes). If we only look at recent blocks (>100000), we still get nine cases of this happening.
