Observations

1. An "invasion" simulation starting with 10 threshold validators at 55 and 1 at 300, very quickly results in the 300 taking over; and indeed the population threshold climbing higher and higher.
2. Similar to the group of validators with threshold ~55, the population after invasion is very inefficient at consensus, with long unused branches.

Evidence

1. The tables below keep track of the mean/standard deviation and the full population descriptions. We see the mean rise very quickly.
   1. thresholdDist.csv
   2. thresholdDist_full.csv
   3. 
2. See the visualizations below.
   1. 
   2. blockDAG_highThreshold.pdf

Conclusions, Speculations and Future Work

1. Even though it seems to be good for the network to have a small number of validators with high thresholds (c.f.