[tyme]

The bill was introduced Friday, so the text is not yet on thomas.loc.gov.
http://thomas.loc.gov/cgi-bin/bdquery/z?d108:h.r.05293:
http://www.fairvote.org/irv/jacksonirvbill.htm

Everyone needs to be aware of the fatal problems with IRV and the existence of better alternatives. Approval voting can be implemented without shifting to ranked ballots, and Condorcet (pairwise) voting uses the same ranked ballots as does IRV. IRV has serious deficiencies when compared with either Approval or Condorcet voting systems.

Condorcet: http://www.electionmethods.org/CondorcetEx.htm
Approval: http://www.electionmethods.org/Approval.htm

IRV is an extremely poor choice. That other countries use it is irrelevant. It is being pushed by various groups without decent discussion of its weaknesses.
IRV fails an entire array of theoretical voting fairness tests:
http://www.electionmethods.org/evaluation.htm

Responses to IRV advocates:
http://www.electionmethods.org/IRVproblems.htm
http://www.electionmethods.org/IRVing.htm
http://www.electionmethods.org/IRVing2.htm

IRV is a ranked voting system. Candidates are ranked from first to last, and the important information content in ranked voting systems is the relative preference of one candidate over another. Sadly, IRV largely ignores these relative preferences, concentrating instead on the first choices, and only moving to second and subsequent choices if no candidate has a majority of first-choice votes. This is dangerously similar to our current first-past-the-post voting system, and does nothing to address the vast array of problems with our current system.

Specifically, under IRV:

If a few voters vote Democrat(1), Libertarian(2) rather than Libertarian(1), Democrat(2), that can illogically shift the winner from Democrat to Libertarian. Only the opposite shift should be possible given that stated change.
As a result, a majority that prefers candidate A over candidate B may be put into a situation where some of them have to rank B over A in order to ensure that A will win.
Voting choices cannot be aggregated. Either the entire ballots must be transmitted to a central location, or the first-choice counts must be retransmitted in each round.


The main advantages of Condorcet, aside from that it passes the above tests, are that it prevents people from gaining anything if they vote in a way other than their true preference ranking. IRV does not. The current system does not.

Those who champion IRV simply don't understand its problems, or don't know that there are alternatives.

A change in voting mechanisms needs to be carefully thought out. Any organization such as fairvote.org that is pushing one voting system and marginalizing all others is either stupid or has an hidden agenda. Fairvote.org mentions some perceived weaknesses in Condorcet voting and Approval voting, yet does not discuss weaknesses in IRV itself. If fairvote.org compared IRV to alternative voting schemes, IRV's deficiencies would be glaringly obvious.

The choice of voting mechanism is a serious decision. Separate bills for each voting method should be introduced, and Congress should debate the merits and weaknesses of each system. Congress and the people need to choose, and fairvote.org is not fairly presenting all the options.

Please contact your representatives and urge them not to consider IRV (h.r. 5293) without considering other options, such as Approval and Condorcet voting.

If you need another reason to be suspicious of this bill, it was introduced by Jesse Jackson.

[Jim March]

Ah. Wow. The first link alone was an eye-opener. I gotta study this in more detail.

If this is true...whoa!

[reagansquad]

Better than what we've got? Yes or no...? I'd say yes.

[tyme]

No, it's not. check out the evaluation at the above link (also referenced three sentences below). Our current system is called "plurality" in that list, and in other literature it's also called "first-past-the-post" voting.

IRV fails criteria that even the Plurality voting system meets.

Again, read section titled "MC" (near the top) at http://www.electionmethods.org/evaluation.htm. Follow the examples all the way through.

I really do not want a voting system like that being used to choose the president, or anyone else.

[tyme]

IRV is a logistical nightmare. When counting votes, there is no way to aggregate them. You can aggregate all the first-place votes to determine which candidate to drop, but then you need access to the original ballots to know which secondary candidates to promote to first place.

Every other major voting method allows aggregation. They allow districts to send only aggregate data to the regional centers, regional centers can send aggregate data to the states. With IRV, either all the ballots have to be in one place, or the states have to tell each precinct to drop a candidate and then resend the aggregate first choice counts. The precincts have no idea if that's the proper candidate to drop.

[tyme]

Here's a perl IRV implementation. I think it's programmatically correct.

Input files look like this:

A;B;C
C;B;D
A;B
A;B
B;A
C;A;B
C;A;B;D
C;B;A

A, B, C, and D can be anything. The ranking separator is a colon or semicolon. Anything else is considered to be part of the candidate names.


#!/usr/bin/perl

open(INFI, "<$ARGV[0]") or die("Sorry, bad filename");

keys %names = 65536;
$irv_round=0;
$votecount=0;

while (<INFI>) {
chomp;
$_ =~ s/ *[:;] */:/g;
$names{$_}++;
$votecount++;
}

print "=== Vote Summary (alphabetical) ===\n";
foreach $key (sort keys %names) {
print "$names{$key} : $key\n";
}

while (!$irv_winner) {
seek(INFI, 0, SEEK_SET);
%names = ();
$irv_round++;
while (<INFI>) {
chomp;
$_ =~ s/ *[:;] */:/g;

@curvote = split(/:/);
IRVDROP: for $a (@dropkeys) {
if ($curvote[0] eq $a) {
shift @curvote;
goto IRVSKIP if ($#curvote == -1);
goto IRVDROP;
}
}
$names{$curvote[0]}++;
IRVSKIP:
}

$cancount = keys %names;
print "\n=== IRV round $irv_round ===\n";
print "$votecount votes, $cancount candidates\n";
foreach $key (sort { $names{$b} <=> $names{$a} } keys %names) {
print "$key : $names{$key}\n";
if ($names{$key} > $votecount/2) {
$irv_winner = $key;
}
$dropkey = $key;
}
push(@dropkeys, $dropkey);

if ($irv_winner) {
print "\n=== IRV winner: $irv_winner on round $irv_round ===\n";
break;
} else {
print "dropping @dropkeys in next IRV round\n";
}
}
close(INFI);

[Brett Bellmore]

Well, it's not hard to tell which voting system YOU like. ;)

Strategic voting isn't really a problem in IRV, so far as I can see, since the exact same circumstances which would make it worthwhile to do, make it impossible to know whether you'd help or hurt yourself. As for the data processing, it's not really a problem today, though it was at one time.

The big advantage of IRV is that it's easy to explain. And that's a huge advantage when you're trying to get people to switch voting systems. If people don't understand a change, they oppose it, and rightly so.

[OF]

It looks cool, but jeeze, in FL they can't even fill out the freaking ballots properly for one guy and you guys want them to choose more than one candidate?

I'll bet you that a large percentage, large, of the country doesn't know anything about the opposite candidate. All they know is they love democrats or what-have-you and republicans are evil. You're not going to get people in this country to successfully 'rank' candidates. They're too ignorant. The horse-race is just what they need, and as shown in 2000, maybe more than they can handle.

Me thinks you're over-estimating the electorates ability to walk and chew gum.

- Gabe

[tyme]

It is NOT cool. This example is straight from the page I cited above.

8: A,C
5: B,A
4: C,B
C loses round 1, B wins round two, 9 to 8.

If two people change their minds from A,C,B to C,A,B (the third choices are implicit once you select the first two):
6: A,C
5: B,A
4: C,B
2: C,A
B loses round 1, A wins round two, 11 to 6.

C was never going to win either way. Two people decided they liked C more than they liked A, and what happened? A went from losing to winning. Even though those two voters didn't change their preference for B, and ranked A lower than before.

[OF]

It is NOT cool.I meant the Condorcet looks cool.

- Gabe

[ProGlock]

tyme:

I have heard David Cobb (Green Party) preach on this subject a lot and he seems to be in favor of it, but of course I'm not entirely convinced it really is a better alternative.

What do you think of this idea....keep our current Electoral College system, but get rid of the "winner take all" that happens with our states. For example currently, if a majority of districts in a state vote GOP, the whole state will go GOP.

I say stop the winner take all in the state. Let the Electoral votes stand on their own and then tally them all up.

[Jay Kominek]

What do you think of this idea....keep our current Electoral College system, but get rid of the "winner take all" that happens with our states.
The states have to decide on this individually. Considering that any state which does this is made irrelevent in the Presidential election, no group which has their states interests at heart will push for it. And hopefully no state will go for it.
(Which is why a Californian is pushing for it in Colorado this year.)

[tyme]

Proglock, that's commonly called "proportional representation" by idiot political scientists. Unfortunately, that's a misnomer.

The difference isn't really proportional vs winner-takes-all.
* Proportional representation means one electorate selects multiple winners.
* Winner-takes-all voting means the electorate is split into subsets, each of which selects one winner.

All the voting systems (IRV, Condorcet, Plurality, Approval, Borda...) are, in their usual form, single-winner systems. They can be modified for multiple-winner elections, but why do that?

This is America. We don't vote for parties. We vote for people. Therefore, we need to know who we're voting for. That's simply not possible if the entire electorate is voting for 435 seats of congress, or even if a state electorate is voting for for 35 electors. The only way to ensure people know who they're voting for, and that they're not voting for some relatively unknown party-picked candidate, is to segregate the electorate so that each subset can focus on a few candidates.

[geekWithA.45]

Except for the issue of "faithless electors", the Electoral College PER SE is nothing more than a vote tabulating system based on 17th century technology.

It is the winner takes all of the district aspect that is the actual foundation of the Republic, ensuring an unambiguous outcome within the district, and ensures that the influence of one district is completely firewalled and isolated from another.

PA's a pretty good example of it working, at the state level. When the district is the entire state, as in the Presidential or gubernatorial elections, PA usually is a swing state, as the 60% ish Dem Majorities in the 3 large cities, plus the odd smattering in the other districts are enough to roughly balance against the entire rest of the state, with the outcome always uncertain.

When the district is smaller than the whole state, the influence of the big 3 are entirely contained to their own borders, and the rest of the place swings mostly Red. As a result, PA continues to quietly enjoy a heftier share of Liberty than most of its neighbors, precisely because PHL can't dictate its liberal whim to the rest of the state.


For _MOST_ elections, as long as the district is reasonable, and no wild gerrymandering gesticulations are in play , this fair, representative, and is the way to do it.

For a minute, I was about to exclaim that just about the ONLY election that might make any sense to do anything differently would be the Presidential, until I consulted the Constitution, as I was suddenly uncertain whether the president was elected directly by the People, as the house of Reps is, or was elected by the State, as the Senate was originally.

The Constitution plainly states that it is the STATE that elects the president, via its electors, and leave open the possibility that the electors may be faithless, split the state vote, etc, as a state issue.

Each State shall appoint, in such Manner as the Legislature thereof may direct, a Number of Electors, equal to the whole Number of Senators and Representatives to which the State may be entitled in the Congress: but no Senator or Representative, or Person holding an Office of Trust or Profit under the United States, shall be appointed an Elector.

Jackson's bill, as it stands, doesn't appear to pass a Constitutional sniff test, as each state clearly has the prerogative "as the Legislature therof may direct.", and is therefore an outrage against the states.