Quote from: ATguy on January 11, 2016, 10:51:16 AM

Quote from: Lauda on January 10, 2016, 09:27:50 PM

This is what I was referencing (such blocks might occur):

You need at least two known values to be able to extrapolate unknown values.

This guenie slide uses just one value, 1 MB with 30 seconds as the latest time any node process a block today. Then just says lets extrapolate unknown values 3 MB, 8 MB with function O(n^2).

This is bad science, you need at least two known values like 0.5 MB and 1 MB with corresponding propagation time to all nodes in order to create estimation of the function for any unknown values like 3 MB or 8 MB. Obviously the more know values you have (over 2), the more precise function you can create and the more reliable extreme far unknown values like 8 MB becomes.

But the slide is big fail with just one 1 MB known value and presented function which comes not from fit with at least two known values but just from author willd guess, thus the obvious and eye catching absurd 3 MB and 8 MB propagation time predictions.

The example represents a severe performance bug in the bitcoin core software and should be fixed by software changes that make an inefficient operation efficient or restrict very large transactions if this proves too difficult. There is no reason to limit blocksize because of this.