Improvements to Inactive Ballot tracking, and corrected reporting of undervotes #857
Conversation
artoonie
force-pushed
the
feature/issue-854_inactive-split
branch
2 times, most recently
from
75341aa
to
f293caa
Compare
artoonie
force-pushed
the
feature/issue-854_inactive-split
branch
3 times, most recently
from
08926ab
to
57af2fd
Compare
...rcv/test_data/2013_minneapolis_mayor_scale/2013_minneapolis_mayor_scale_expected_summary.csv
Outdated
Show resolved
Hide resolved
|
I laid out all the open questions in the original issue #854 |
artoonie
force-pushed
the
feature/issue-854_inactive-split
branch
3 times, most recently
from
a26818b
to
723b4b8
Compare
artoonie
changed the title
artoonie
force-pushed
the
feature/issue-854_inactive-split
branch
4 times, most recently
from
52cd0cd
to
a595e69
Compare
artoonie
force-pushed
the
feature/issue-854_inactive-split
branch
from
a595e69
to
55eee98
Compare
artoonie
changed the title
...brightspots/rcv/test_data/2013_minneapolis_mayor/2013_minneapolis_mayor_expected_summary.csv
Outdated
Show resolved
Hide resolved
...brightspots/rcv/test_data/2013_minneapolis_mayor/2013_minneapolis_mayor_expected_summary.csv
Show resolved
Hide resolved
|
note for when this is merged that RCVRC needs to document how this changes the description for ballots in the audit log and potentially the totally unranked ballots in the summary.csv |
| @@ -1172,20 +1174,8 @@ && isCandidateContinuing(cvr.getCurrentRecipientOfVote())) { | |||
| // if this is the last ranking we are out of rankings and must exhaust this cvr | |||
| // determine if the reason is skipping too many ranks, or no continuing candidates | |||
There was a problem hiding this comment.
ah, this part was always so confusing to me! But I think I understand now. The code only gets here if all of their ranked candidates are not continuing. We were moving them to undervote if their ranked candidates were exhausted and they didn't use all their rankings. The code that was removed was using the implicit partially/fully ranked code which might be made moot by the discussion in #854
So basically with this update when you get here you will have your ballot marked exhausted no matter how many rankings you've used. right??
If so, the second line of this comment should be removed
| Inactive Ballots by Exhausted Choices,0,,18,18,,0,18,,1,19,,1,20,,4,24,,5,29,,5,34,,2,36,,15,51,,14,65,,15,80,,11,91,,16,107,,27,134,,21,155,,21,176,,19,195,,28,223,,47,270,,46,316,,89,405,,101,506,,161,667,,100,767,,95,862,,306,1168,,336,1504,,261,1765,,236,2001,,535,2536,,1406,3942,,7827,11769,,0 | ||
| Inactive Ballots by Repeated Rankings,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0,0,,0 | ||
| Inactive Ballots Total,150,,37,187,,6,193,,5,198,,11,209,,10,219,,17,236,,12,248,,17,265,,33,298,,29,327,,33,360,,37,397,,22,419,,40,459,,51,510,,29,539,,25,564,,72,636,,54,690,,56,746,,99,845,,124,969,,270,1239,,142,1381,,121,1502,,509,2011,,370,2381,,458,2839,,300,3139,,733,3872,,1971,5843,,9825,15668,,0 | ||
| Contest Information |
There was a problem hiding this comment.
any idea why this whole file is flagged as new but with actual changes highlighted in darker green????
There was a problem hiding this comment.
Line ending change (must've previously been created before we made all line endings the same).
If you select "Hide Whitespace" in the diff view it goes away:
| @@ -100,4 +100,4 @@ | |||
| "totalNumBallots" : "26", | |||
| "undervotes" : 0 | |||
| } | |||
| } | |||
| } | |||
There was a problem hiding this comment.
A "no EOL" symbol -- removing the trailing \n from the file to line up with most of the other JSONs
| Number to be Elected,1 | ||
| Number of Candidates,1 | ||
| Total Number of Ballots,10 | ||
| Number of Ballots with No Rankings,0 |
There was a problem hiding this comment.
asked @chughes297 and confirmed this change from undervote to exhausted is correct.
|
@artoonie can we change the undervote description in the |
|
All comments addressed, let me know if there's more needed from RCVRC's side |
This was originally designed to handle both #780 and #854. However, #854 has additional questions and has been deferred out of 2.0.
Still, this PR made some improvements that resolved #780 in a nice way, so I'm modifying this PR to just resolve #780 without the "partially ranked" vs "fully ranked" split in #854.
Therefore, this PR has code that is outside of the scope of #780 but makes it easy to accomplish #854 in the future -- things like tracking whether a ballot used all of its known rankings.