From a practical standpoint in terms of conducting an election:
Voting happens only once; no “multiple rounds” of voting.
- It is easy to understand and to do, and to educate the voters how to do it.
Each voter gets a single ballot, and fills-in the space in the appropriate preference column for each candidate, which effectively numbers the candidates from first preference onward.
Most
1← Preferred →
2
3Least
4Alpha
[party 1]█[][][]Beta
[party 3][][]█[]Gamma
Independent[]█[][]Delta
[party 2][]█[][]In the event it is a Condorcet-MMPR (general) election, the voter would also fill-in the space in the party-list section of the ballot, to signify his or her party-preference.
Most
1← Preferred →
2
3Least
4Preferred
PartyAlpha
[party 1]█[][][][party 1][]Beta
[party 3][][]█[][party 2]█Gamma
Independent[]█[][][party 3][]Delta
[party 2][]█[][]
Counting happens only once at each poll:
No multiple rounds of tallying the vote nor of reallocating of ballots as candidates are “eliminated,” as would occur with IRV, for instance (or STV) — with Condorcet/Ranked-Pairs we do NOT eliminate candidates, but apply all preferences holistically.
Candidate
PairsB more-
preferredNo-
PreferenceA more-
preferredTotalA: Alpha
vs B: Beta27667100A: Alpha
vs B: Gamma48052100A: Alpha
vs B: Delta22870100A: Beta
vs B: Gamma60139100A: Beta
vs B: Delta80218100A: Gamma
vs B: Delta22078100- Depending on the number of candidates (see Practical Issues ), the ranked-pairs tally can be done manually; it is more labour intensive than FPTP, and to count a poll on election night will take longer.
- With more candidates we would need to count by means of an optical/marksense reader, or some other computerized means, but such a count would be at least as fast, if not faster, than at present.
One evaluation round:
- The count for each poll is communicated to the constituency Electoral Officer on election night to be aggregated with counts from other polls into a count; and
Once all polls have reported, the constituency Electoral Officer determines the ranking of the candidates, and thus the winner (single representation) or winners (multiple representation).
Majority
CandidateMinority
CandidateAffirmAccrued ResultDelta
82→Beta
20✔Delta → BetaGamma
78→Delta
22✔Gamma → Delta →
BetaAlpha
78→Delta
30✔Gamma → Delta →
Beta
AND
Alpha → Delta →
BetaAlpha
73→Beta
33✔Gamma → Delta →
Beta
AND
Alpha → Delta →
BetaGamma
61→Beta
40✔Gamma → Delta →
Beta
AND
Alpha → Delta →
BetaAlpha
52→Gamma
48✔Alpha → Gamma →
Delta → Beta
The system is easy for voters to learn, to understand, and to do. It is relatively easy for election officials as well.
It can be:
- A plug-in replacement for single-member systems (Single-member FPTP), or an IRV system, in which case we can keep the same constituencies and continue to elect only one person per constituency, chosen as the most-preferred candidate per the Condorcet/ranked-pairs analysis; or
- A plug-in replacement for an existing multiple-member plurality (Multiple FPTP, or STV system, etc.) where we elect n candidates for a given constituency — we keep the same constituencies and select the n most-preferred candidates from the ranked-pairs analysis; or
- A plug-in replacement for an existing MMPR system upgraded to a Condorcet-MMPR system — instead of electing constituency representatives by FPTP, as before, we do so by Condorcet/ranked-pairs; or
-
A single member system upgraded to a Ranked-Pairs MMPR system (we either keep the same constituencies and increase the size of the legislature to accommodate the party-list seats, or we reduce the number of constituencies according to the nominal number of party-list seats) — for the given constituencies elect the most-preferred candidate per the corresponding ranked-pairs analysis.
For general elections with MMPR we also capture voters’ party preferences, and once the standings of all constituencies are determined, top-up the constituency representatives from party-lists to achieve proportional representation for the legislature as a whole.
Next: Practical Issues