Correcting the error in getnetworkhashrateps

The discussion revolves around the application and understanding of the Erlang distribution in contrast to the Poisson distribution within the context of blockchain network analysis, particularly focusing on how blocks are timed and counted. The Erlang distribution is highlighted as more appropriate due to the method of using a fixed number of blocks to measure time, rather than a fixed period. This approach challenges the traditional application of the Poisson distribution which assumes a fixed interval of time for its calculations to be accurate.

An important aspect brought up is the concept of selecting a random point in time to determine the interval to the previous and next block, which inherently assumes a memorylessness property. This property suggests that the time until an event (in this case, the discovery of a block) is independent of any past events. However, the selection of a timestamp that coincides with a block introduces a discrepancy because it does not account for the time before and after this block accurately, hence suggesting an adjustment formula (N-1)/N to correct for the additional block included in the calculation.

This adjustment is necessitated by the observation that when a specific timestamp is selected—especially one that aligns with the discovery of a block—it inadvertently counts one block too many for the analysis to remain accurate under the assumptions of the Poisson distribution. Therefore, to reconcile this issue and maintain the integrity of the analysis, an adjustment is proposed to subtract the excess block from the overall count, ensuring the model remains reflective of the true dynamics at play within the blockchain's operational framework.