The original Satoshi paper uses the Poisson distribution as an approximation to model the bitcoin mining process. The more appropriate model might be negative

The original Satoshi paper uses the Poisson distribution as an approximation to model the bitcoin mining process. The more appropriate model might be negative binomial, because you want to know how many successes you get until a specified number of failures occur. I don't doubt the Poisson distribution would model the data fairly well, but it's really hard to say that a real process is Poisson distributed by definition.