You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am not convinced the number of points is a good way to parameterise this. Tn that case, the resolution depends on the number of transactions in the mempool. E.g. if the user asks for 10 points, in a mempool with a 200B transaction, you get a sample every 20B (needless accuracy) whereas in a mempool with 100MB worth of transactions, you get a sample every 10MB (way too inaccurate).
Maybe a more interesting would be to specify the sampling interval and a max limit. E.g. sample every 200kB and max 50 samples to cover top 10 blocks. Probably more intuitive that way.
As a bit more out there idea (not for this PR), maybe the sampling should not be uniform but logarithmic. I can imagine the feerate distribution close to the top of the mempool is more interesting than 10s or 100s of MB down the line.
This will not be a breaking change it will only change the number of points being retuned to better capture the mempool's feerate curve
The text was updated successfully, but these errors were encountered:
This will not be a breaking change it will only change the number of points being retuned to better capture the mempool's feerate curve
The text was updated successfully, but these errors were encountered: