I understand, but every probability function has some "noise" characteristics, and can be described. Extremely long run without blocks should be less probable.
Though, I know this is not pseudorandom, but true random. But still - if you statistically count these runs, the most extreme ones should be the rarest, shoudn't they?
Statistics undergrad here.
I've pasted your figures into R, and here's the resulting P-P plot, showing no significant deviation from exponential distribution.
http://www.pixentral.com/pics/1siXeqeLltBWny5zoatjq4BM2A0xwV1.png