I run a distributed software making rounds. Each round has a 35% probability of success. I've been monitoring it for quite some time, and I have the intuition that the successes are happening in "packs", example 3 successes in a row then 7 fails, etc etc...
I already have a significant amount of repetitions. What metric can I use to verify if the successes tends to make "packs" or if it was a wrong impression?