Milestone selection algorithm

[gal.rogozinski]

@luigi.vigneri

According to the above properties, a randomly chosen tip will belong to the heaviest branch whp (see the dark green tip of Figure 1). The choice of the second tip follows the same reasoning, hence it is chosen randomly.

What I am thinking, since anyhows we have to do a walk to count transactions, why not use Local Modifiers walk to choose the tips for the milestone

This is for the following reasons:

1. It is less gameable than (almost) URTS on a subset
2. In order to implement the secret transaction idea, we have to do a walk anyhows…
3. Using LM will increase the likelihood of us using a tip straight from the get go that satisfies \Gamma

I guess the walk that starts from a secret checkpoint can have lower randomness and a walk that starts from a milestone can have higher randomness (just intuition)