This is a simulation story. We should make sure that the experiment does indeed sweep over the specified parameters:
The PSuU Experiment is Required to sweep over through the following parameters:
[x] Phase Durations
[ ] Random Start Times or TS to be used
The PSuU Experiment is Not Required to sweep over through the following parameters:
[x] No sweeping over rewards, slashes, commitment bond, gas prices, thresholds, etc
[x] Initial Agent count is to be set to 10. Initial Stake is to be set high enough so that we don't have drop-outs. Initial Balance is to be set high bc we are not interested in its implications. eg. 100_000 ETH per agent.
Notes
Empirical Data
[ ] Fixed Censorship Builders (Should be in nb)
[ ] Data used: From Dencun -> Today
[ ] Random Start times from subset of Dencun -> Today. Currently, we set a start_time. Unless we want to use the same subset on every run, and unless we want to sweep over different start times (which would get really heavy on complexity), we could just pick a random start_time for each run.
Old
[ ] Using reasonably large number of agents, but no need to go crazy. Biggest risk is that low amount of agents results in them getting slashed, dropping out, without dropping back in, distorting results. All trajectories are going to have the same initial state for the agents (eg. # of agents, properties, etc)
This is a simulation story. We should make sure that the experiment does indeed sweep over the specified parameters:
The PSuU Experiment is Required to sweep over through the following parameters:
[x] Phase Durations
[ ] Random Start Times or TS to be used
The PSuU Experiment is Not Required to sweep over through the following parameters:
[x] No sweeping over rewards, slashes, commitment bond, gas prices, thresholds, etc
[x] Initial Agent count is to be set to 10. Initial Stake is to be set high enough so that we don't have drop-outs. Initial Balance is to be set high bc we are not interested in its implications. eg. 100_000 ETH per agent.
Notes
Empirical Data
Old