You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Although we are not required to audit STV contests (i.e. multi-winner preferential contests), we need to ensure that colorado-rla does not break when they are uploaded. This requires a bit of thought. Probably we want three types of contests: plurality, IRV and STV.
Probably:
when parsing, a contest with ranks and multiple winners should be STV,
STV contests should be always NOT_AUDITABLE
STV contests cannot be targeted for audit
STV winners should have a clear placeholder (e.g. UNKNOWN or "") - note that the default contest counter will get the wrong winner, just as it does for IRV.
STV estimatedSampleSizes and optimisticSampleSizes should match the values for other NOT_AUDITABLE contests. (I have a vague recollection that this is 0, but we should check.)
Think about whether we could easily get a count of 'other discrepancy' in the event that an STV contest happens to be on an audited ballot.
Consider whether all this is best achieved with an STVComparisonAudit class.
The text was updated successfully, but these errors were encountered:
This will probably also need to update ReportRows::genSumResultsReport, just to make sure the tallies computed by countContest are not output for STV (where they will be wrong).
Although we are not required to audit STV contests (i.e. multi-winner preferential contests), we need to ensure that colorado-rla does not break when they are uploaded. This requires a bit of thought. Probably we want three types of contests: plurality, IRV and STV.
Probably:
Think about whether we could easily get a count of 'other discrepancy' in the event that an STV contest happens to be on an audited ballot.
Consider whether all this is best achieved with an STVComparisonAudit class.
The text was updated successfully, but these errors were encountered: