Currently, for some protocols, each validator is randomly assigned an initial starting estimate. For example, in Binary consensus, each validator is randomly assigned an initial estimate. Ideally, we would be able to control these initial estimates.
Proposed Implementation
Protocols can store a dictionary from validator name => initial estimate, in the protocols that it makes sense. Some function can populate this dictionary with initial estimates.
If some initial messages are specific, the validators should use these in their first bet. Otherwise, they can choose a random estimate, as they do currently.
Issue
Currently, for some protocols, each validator is randomly assigned an initial starting estimate. For example, in Binary consensus, each validator is randomly assigned an initial estimate. Ideally, we would be able to control these initial estimates.
Proposed Implementation
Protocols can store a dictionary from validator name => initial estimate, in the protocols that it makes sense. Some function can populate this dictionary with initial estimates.
If some initial messages are specific, the validators should use these in their first bet. Otherwise, they can choose a random estimate, as they do currently.