[EM] IRV vs Plurality

[Abd ul-Rahman Lomax]

At 11:22 AM 1/15/2010, robert bristow-johnson wrote:
(about a voting security expert)
>you are in the rabid anti-IRV party.

Robert, your slip is showing.

Experts in various fields tend to be strongly against IRV. Political 
activists who are working for IRV tend to see strong opposition as 
"rabid." You want to see rabid opposition, you'll have to look 
elsewhere, though.

I'm thinking over the opposition material. It might look rabid to 
someone who isn't aware of the problems, and there are some opponents 
who can lose perspective, but the more I learned about IRV, the more 
I realized what damage it can do. It's not as simple as being a 
terrible system. In some ways, it isn't terrible, in terms of 
results. But there are, indeed, serious problems with our voting 
systems, and if a "reform" is implemented, with high hopes -- 
fantasies, really -- and at high cost, it can impede the progress of 
true reform for decades.

That's why I recommend starting with the simplest and cheapest 
reforms. As simple and as cheap as counting all the votes. And this 
aligns with the interests of the voting security experts. They also 
want us to count all the votes!

Just that one change, with a corresponding change in instructions, 
turns Plurality into a quite sophisticated voting system. No change 
in voting equipment (it must already be able to handle multiple votes 
in one race because of plurality-at-large elections, plus there are 
other ways to arrange the ballot even if this approach is for some 
reason impossible).

Is that system ideal? I think not, but it's an improvement, and I've 
never seen a cogent argument that tossing "overvotes" was better than 
keeping them. There are a series of knee-jerk arguments we see. 
Supposedly it violates one-person, one-vote, but it's easy to show otherwise.

And that improvement, which will *at worst* do nothing, it defaults 
to Plurality, but we can be sure that not all voters will bullet 
vote. In particular, as to the big concern about the spoiler effect, 
the supporter of a minor no-hope candidate gains an option and no 
longer must face a Hobson's choice: betray the favorite or lose 
voting power. And when we do instant runoff approval voting, the 
voter can vote that first preference, and if a majority agree, the 
first preference wins.

(I won't go into all the various options and tweaks that would 
further improve IRAV, because the point here is that a no-brainer 
improvement that costs nothing should be first to be considered, not 
last. And if we are going to do IRV, we must ask, why not allow 
voters who want to do so to equal-rank? Why toss those votes? What 
harm would be done? And, I can tell you, IRV gets much better if 
multiple approvals are allowed. But ... there are even better 
methods, and why use the complicated elimination system of IRV, which 
requires central counting or centrally coordinated counting, where an 
error in any precinct can require all other precincts to recount. Is 
it "rabid" to note these severe problems?)

There is a particular problem with activists who are trying to change 
the status quo. They think they have a better idea, but anyone who 
strongly points out the problems with this idea they will then 
identify as a rabid supporter of the status quo. Robert, you have 
done this, and you have pushed an expert into your imagined "rabid 
anti-IRV party." In this, you have quite strongly allied yourself 
with FairVote, which has sought to so identify all opponents of IRV, 
dismissing Warren Smith and others as simply biased against IRV, 
often presumed to be because they are "pro-Range," when, in fact, 
Smith is a mathematician and, while he got all hot and bothered some 
years back over his realization that Range could have some very 
salutary effects, and somewhat thought of the situation as an 
emergency, he never allowed himself to leave behind his balance, and 
he openly acknowledges problems with Range and studies them.

I don't agree with Smith's overall political analysis. But, after 
all, that's not his expertise. The expert you have attacked is not an 
expert -- not yet! -- on voting systems, but on voting security. 
However, she's learning. Are you? It seems you have learned 
something. You started as an IRV supporter. You listened to enough of 
the criticism to get that it wasn't without basis.

Once upon a time, I'd have said, okay, IRV has problems, but it's 
still better than plurality, so I'd support a measure to replace 
plurality with IRV. I am no longer so confident of that, and it might 
depend on details, among them the specific application. Take Takoma 
Park, Maryland, which was an early IRV adopter in the current fad. 
Why do they have IRV? It makes no sense there. The majority of 
elections are unopposed, common in small towns, where it can actually 
be difficult to get anyone to run, sometimes. A few elections have 
two candidates. And a very few have more than that but are easily 
decided in the first round, in reality. The elections are 
nonpartisan, and we know that in nonpartisan elections in 
environments like that, there is no difference between the results of 
IRV and the results of plurality. So .... if I lived in Takoma Park, 
would I support IRV?

It would depend on another factor, my identity. If I were Rob Richie, 
I'd support IRV there, because it would further my national campaign. 
And you can guess where Rob Richie lives.... IRV was harmless in 
Takoma Park because they basically get the result in the first round, 
at least most of the time, and the results are, after all, the same 
as plurality. So why not help out a prominent town resident?

But where there are elections with 23 candidates on the ballot, IRV 
gets very expensive. And the method breaks down and starts to show 
serious pathologies. Does it produce better results than Plurality? 
No. The same results. Show me an exception in a nonpartisan election 
and we can discuss this, there may actually have been one, but the 
rarity proves the point. And IRV wasn't replacing plurality in San 
Francisco, it was replacing top-two runoff, which, for reasons which 
are not necessarily obvious to people who haven't deeply studied 
voting system theory and experience, is a much better method than 
IRV. It was a step backwards. At great expense, though in one way it 
may have been worth it.

It gave us, quickly, a large body of data allowing to study the 
actual performance of IRV. Who is paying attention to that data? Only 
the people whom you might call the "rabid anti-IRV party." Has it 
occurred to you, Robert, that IRV might be much worse than you've 
realized and that these people might actually be onto something?

Yes, there are some anti-IRV activists who are, in fact, attempting 
to preserve Plurality, it seems, and who may have other suspect 
motives; the prominent group in Minnesota uses arguments that are 
sometimes pretty shaky. There is also a truly excellent video 
producer on YouTube who is astonishingly effective in calmly pointing 
out the problems with IRV, and I haven't seen him proposing 
alternatives, so is he a "Plurality-pusher"? I don't know. I don't 
know who he is, but, boy, if I had a political campaign, I'd want him 
on my side. He's not "rabid," and that's why he's so effective. He 
calmly leads his audience into a clear realization of the problems.

You really should study the history of the relationship between 
FairVote and voting systems theorists. The Center for Range Voting 
people, basically Warren Smith and Jan Kok, with my assistance, 
attempted to find ways to cooperate with FairVote. It was impossible, 
Richie didn't give a fig about broader consensus, he's a professional 
political activist and he has one goal: victory. Victory for what? 
Whatever he's pushing, and he's not about to dilute it or weaken it 
with compromise. To be fair to Rob, he's just doing what many other 
political activists do, it's the norm, in fact. When you decide to 
support a candidate, you don't back up and allow yourself to have 
doubts, you persist and do  your best at least until the election is 
over. It's like being a lawyer for the prosecution or defense. If the 
defense lawyer starts to have doubts about the guilt of the client, 
he or she will nevertheless do his or her best to convince the jury 
that the client is innocent.

And there is even a value to this. If the decisions are made in an 
environment where there is deliberation, as happens with a jury. They 
talk it over, review the evidence, and (when the system is working) 
the evidence is all laid out, with expert testimony as needed. But 
this is in a situation where the question is simple: Yes (Guilty) or 
No (Not Guilty) or No Decision (Hung Jury), and with serious issues 
(criminal cases), unanimity is required. (With civil issues, it's 
simple majority, but no decision can be made by a jury without a 
majority consent. They *never* would use plurality or IRV. But they 
might use approval voting in their process, or even deeper Range, to 
poll the jury, it's a great idea. The final result would always be 
ratified by a majority, so there is absolutely no damage from the 
asserted pathologies caused by supposed strategic voting.)

But there is a better way, in fact; the adversarial technique is part 
of the problem with our social decision-making system. It tends to 
drown out the still, small voice of knowledge and wisdom, in favor of 
heat and fervor.

Robert, I'm suggesting to you a fast path. Drop your attachment to 
what you think you know, much of it is shallow and incomplete. I'm 
not suggesting that you rush to the opposite side and accept 
everything you are being told. Be skeptical, ask for proof or 
evidence or clear explanation.

A long time ago, I realized that if I met some perfect expert who 
knew everything, and whose analysis was deep and accurate, I would 
surely disagree with him or her on many issues....

That doesn't mean that I'm wrong on any particular issue, and there 
is no such expert. But it does mean that I'm surely wrong on some 
things, and how am I going to correct myself if I automatically 
believe that I'm right?

So I discuss stuff, and, for sure, I've stuck my foot in my mouth 
many times. Discuss stuff with experts when you aren't an expert, the 
fastest way to learn is, in fact, to taste your foot. And learn from 
it. Don't get stuck in personality conflicts, they will distract you 
and cause you to harden your positions. After all, who wants to admit 
that the arrogant jerk was right?

Sometimes that is exactly the best thing to do, and it's admirable. 
It doesn't make an ass into a nice guy, or anything but right on that 
point, for even a stopped clock is right twice a day, and we could 
say that sometimes the stopped clock is, at those times, more correct 
than the best and usually most accurate clock, because the latter is 
almost always off by some value, so it is almost never fully correct! 
But it can be made more and more correct, the average error is far 
lower than for the stopped clock. I prefer not to be a stopped clock 
until I'm dead. And maybe not even then, we'll see, eh?