[EM] All combinataions of weaker-preferringness & conditionality. Eric Gore's Deterministic RP.
Michael Ossipoff
email9648742 at gmail.com
Sun Dec 15 06:22:07 PST 2013
Regarding the 3 paragraphs of HDRP, each paragraph could have, or not
have, the conditionality clause,&/ or the weak-preferring clause. I'd
like to post, here, a naming system for those combinations.
It has occurred to me that there _is_ something different about the
3rd paagreaph, that qualifies it more for both clauses. RP's
definition calls for discarding defeats that contradict not-discarded
stronger ones. To have either clause in paragraph 1 &/or 2 would be to
not really be doing RP.
So I suggest that those clauses, if used, should be confined to
paragraph 3, if the method is to really be RP.
I don't know that the result would be bad if the clauses were using in
ppg 1 &/or 2. Maybe it would be ok, but it wouldn't really be RP, and
so I don't suggest it.
As for ppg 3, RP doesn't say anything about discarding every defeat
that contradicts only equal defeats. It doesn't say to make 3-way ties
willy-nilly. For that reason, I'd prefer having the two clauses in
paragraph 3. That's where both of the clauses are permissible for RP,
and desirable for RP.
If I just say "HDRP", that will refer to HDRP with both clauses in
paragraph 3, and neither in the other paragrahs.
If none of the paragraphs have either clause, that's Unconditional HDRP.
If none of the paragraphs have the weaker-preferring clause, but
paragraph 3 has the conditionality clause, that's Conditional HDRP.
Other than that, I name the other combinations as follows:
HDRP Conditional [#s of the paragraphs that have the conditionality
clause] weaker-preferrring [numbers of the paragraphs that have the
weaker-preferring clause].
So Unconditional HDRP, as defined above, could be written:
HDRP conditional none, weak-preferring none.
Conditional HDRP, as defined above, could be written:
HDRP conditional3, weak-preferring none
HDRP (withoujt any other designation) could be written:
HDRP conditional3 weak-prefferring3.
Here is a deterministic Ranked-Pairs that is defined at Eric Gore's website:
[quote]
Starting with the strongest defeat, consider each defeat in sequence
with previously kept defeats, if any. If two or more defeats are
equivalent, those defeats are considered together with previously kept
defeats, if any. If any defeat under consideration, which has not yet
been kept, is a part of a cycle, it is rejected. If any defeat under
consideration, which has not yet been kept, is not a part of
[/quote]
A time-independent way of writing that would be:
Greater & Equal Deterministic RP (GEDRP):
A defeat,D, is a discarded defeat if it contradicts a set of defeats
that contains 1 or more not-discarded defeats stronger than D, and
doesn't contain any defeats weaker than D. .
[end of GEDRP definition]
CIVS-RP, in that naming-system, would be Greater Deterministic RP (GDRP)
Greater Deterministic Ranked-Pairs:
A defeat is a discarded defeat if it contradicts a set of
not-discarded stronger defeats.
[end of GDRP definition]
So here are some kinds of Ranked-Pairs
Tideman's Ranked-Pairs (margins, no equal or truncated ranking)
MAM
Deterministic RP
...Greater Deteministic Ranked-Pairs (GDRP)
...Greater & Equal Deterministic Ranked Pairs (GEDRP)
...Hierarchial Deterministic Ranked Pairs versions
...(HDRP versions)
......HDRP (both clauses (conditional & weaker-preferrng) only in ppg 3)
......Conditional HDDP (only conditional clause in ppg 3)
General HDRP naming:
HDRP conditional [ppgs w/ conditional clause] weak-preferring [ppgs w/
weak-preferring clause]
MMV, as defined at Prabhakar's website, is:
A defeat, D, is a discarded defeat if it contradicts a set of defeats containing
1 or more defeats equal to D, and/or one or more not-discarded defeats
equal to D, but contains no defeats weaker than D.
[end of MMV as defined at Prabhakar's website]
Of the detrministic RP versions, I prefer the HDRP version, HDRP in particular.
(That's the HDRP version with the conditional and weaker-preferring
clauses in paragraph 3 only).
Michael Ossipoff
More information about the Election-Methods
mailing list