Q. OK, OK, I concede IRV is a worse voting system than Range Voting, but it has the Reform-Momentum right now. So you by promoting range voting are doing us all a disservice and being tactically unwise by splitting voting-reform forces.

(Executive summary)

A. Exactly wrong: we shall demonstrate that Range Voting is actually unifying reform forces that would otherwise remain divided.

It unfortunately is correct there has been division among voting-system reform advocates, with the "Center for Voting and Democracy" pushing IRV, the "Center for Approval Voting", the "Center for Range Voting" (US) and condorcet.org pushing Condorcet. And you are right that this internal dispute could paralyse voting reform.

But there is one system that is better than all the others, and which the US third parties (who will have to be the troop-supplier on this one) can unify behind: range voting. Once education occurs, there is no split anymore. Because:

  1. Range voting experimentally generates much greater support for third parties than either approval or plurality voting.
  2. Range voting is usable on today's voting machines – while IRV and Condorcet are not – which means it has a lot more chance of getting adopted.
  3. IRV leads to self-reinforcing 2-party domination, just like plurality voting – meanwhile range voting (probably) doesn't (and plurality with a genuine separate runoff, definitely doesn't).
  4. The best Condorcet methods are far more complicated than range voting, quite plausibly also lead to 2-party domination, and there is no consensus on which of the many of them is the best.
Third parties are not complete idiots. Given these facts, they are going to unify behind range, not the others. That means only range voting can get unified dedicated core support. So only it can achieve reform.

OK? It is not a matter of Range Voting splitting reform forces and preventing reform. It is exactly the opposite: range voting engenders unification of those forces allowing reform for the first time.

So once that is realized, support for Condorcet, IRV, and approval is gone. We are slowly educating third parties, and as they get educated, they switch to range and do not come back. IRV is in fact suicidal for third parties since it just leads to 2-party domination e.g. in all countries that've tried it.

Moral: The "split" is an illusion. There is no split, there is only a temporary illusion caused by a slowdown in education. If you support IRV then you are merely prolonging that bad and deluded state of affairs by slowing education and fostering delusions. You may also find it interesting that the (first) inventor of approval voting, Ottewell, is now endorsing range voting and the CRV as superior.

Think of IRV as snake oil and range voting as a genuine cure. Benefit for society is not got by pushing snake oil, even if the polls say it is temporarily more popular than the genuine cure by an incredibly tiny margin. IRV is like a stock bubble – may look temporarily like a good investment, but it isn't!

Now you might object that maybe the above has placed too much emphasis on third parties. We need support from the top two parties. Good point. So consider our main push – to get Range Voting in the Iowa 2008 caucuses. Suppose we were to instead push IRV in Iowa 2008. That would be a tactically worse idea with far less chance than Range of succeeding and which would yield a less-good result even if it did succeed:

  1. The top Democrats and Republicans who would make this change, see an upside with Range – better presidential choice for them so they can win – and a downside. What is the downside?
  2. Their downside: risk of major election screwup ala the San Francisco IRV experience in 2004 or ala Florida 2000 (near tie between Gore & Bush – IRV offers many tie possibilities, one per round, and hence greatly increases the risk of this sort of nightmare, versus plurality voting). Also, the Democrats & Republicans risk looking stupid with IRV if mostly plurality-style "truncated ballots" are cast (if they allow truncation) or furious voters (if they instead demand full rank ordering). Result: ridicule for them and their party. Ultra-bad news.
  3. Unlike range, IRV cannot be done via just having precincts compute totals and send them in. That means with IRV, precincts have to compute first round totals. Then send them in. Then wait until every precinct has reported. Then report the first round results. Then precincts compute second round totals. Then send them in. Then wait until every precinct has reported. Then report the second round results. Then precincts compute third round totals... Then precincts compute 10th round totals. Then send them in. Then wait until every precinct has reported. Then report the 10th round results. If ever an error is found, that could force everyone, statewide, to go back to the beginning to the massive annoyance of all. This is an election administration nightmare if it is conducted manually by precinct as currently done in Iowa. If it were conducted by computer, then it could be done, but then you'd have to have every computer connected in a statewide network (another nightmare) or you'd have to ship all the ballots to one Iowa central location, where they all would be counted centrally (another nightmare and a complete change in procedure from the current system – i.e. just asking for trouble).
  4. Are the top Ds and Rs idiots? Why risk ridicule? With range, they get the same upside (in fact better upside than IRV – higher quality election results on average as proven by extensive computer simulation studies) but no downside (range is in fact superior to plurality voting in the senses of having less chance of a tie and is far simpler to total than IRV hence no chance of San Francisco style screwups).
  5. The third parties will refuse to help with voting system reform efforts (except perhaps in a confused and disunified manner and they'll be angry later when they find out they were scammed) if those efforts push IRV in Iowa. But they will help reformers if the push is for Range Voting. Because range benefits third parties, but IRV doesn't.

Another advantage for reformers is that range leads to less backslide risk. (Many US cities including NY City adopted IRV voting, especially in the 1930s, but almost all of them backslid and converted back to plurality voting. In contrast I am unaware of any entity that, once it adopted range voting, backslid. Ever.)

One backsliding rationale (used by the IEEE as justification for their own backslide) is that with IRV and Approval Voting, a lot of voters want to vote plurality style ("truncated ranking"). If you force them to provide full rankings then the system is unpopular and complicated for the voter, and highly likely to lead to invalid ballots, and then you are exactly wrong in saying that IRV is simpler for the voter. Range is simpler: no permutation constraint – linear – additive – runs on today's voting machines – and voters can leave candidate scores blank if they feel ignorant about that candidate. In IRV, blanks are not possible except via truncated ballots (which oftentimes are disallowed) and this gives the wrong effect anyhow. (For example, even in IRV with truncation, a voter simply cannot express the idea that Gandhi is best, Hitler is worst, and he feels ignorant about Perot and Anderson.) A small subset of Australia has allowed IRV ballot-truncation and the result in Kuring-gai was 63% plurality-style IRV voting. That behavior provides good rationale to backsliding advocates to go back to plurality! Meanwhile plurality-style range ballots experimentally are rare – only 19% even in the extremely highly plurality-genic 2004 US presidential election (in which the top two candidates got over 99% of the plurality votes – this is from our 2004 range voting presidential election study #82 here ) – so there will be no such no backsliding rationale with range voting.

OK? So I hope you can see this pro-IRV anti-range argument – good as it might initially sound – was deluded. Range is better tactically, less likely to be a reform that merely backslides, and range is better quality-wise and range is less confusing (not more) for both the voter and for the election officials, and range is much less likely than IRV (and also less likely than plurality) to lead to nightmares for election and party officials.

And just how much "momentum" has IRV got that we risk sacrificing, anyhow? Not a whole heck of a lot. Most US citizens don't even know what IRV is. Not a single US state has adopted it. A great number of US cities have abandoned STV schemes including New York City and Cincinnati. The "momentum" should be interpreted as "in favor of improving the voting system" not "about IRV." That is the real direction of the momentum. if so, then nothing is lost by advocating a superior system – range voting.

A zen anecdote: "Global cooling sounds like a great idea, but we need to unify behind global warming, because it has all the momentum!"

Return to main page