Well, I did that in a consistent statistically powerful and apparently correctly executed test some time ago and the result was that TDS resurrection is not random and independent events. This means that probability of each individual event may be close to 0.25 but cumulative probability of multiple events cannot be calculated simple my multiplication of these probabilities. So, for example, chance to resurrect twice in a row is not 0.25^2, maybe something else, IDK what. Also, since this is pRNG, the probabilities should be probably dynamically calculated. It is most likely related to theory of multidimensional Bayesian probability although not very clear exactly for me. It is also quite a complex case of multidimensional algebra.
There was some extensive discussion of GoW-related statistics and this post has also links to test results.
And while streak should normally happen in any type of random number sequence, since TDS resurrection is not random and independent, these streaks are not regular random streak but indeed something different. Seems like Sirrian does not know what it might be. This game is quite complex.