Blogging Bayport Alameda

November 24, 2010

Running the numbers

Filed under: Alameda, City Council, Election — Tags: , — Lauren Do @ 6:09 am

One last meaty blog post before the holiday weekend.

It is unsurprising given the results of the election and the shocking win of Jean Quan in Oakland, that some Alamedans that appeared to have not supported the ultimate victor in the Mayor’s race are now clamoring for an adoption of Ranked Choice Voting in order to better capture the true sentiment of who Alamedans want in their elected office.

The unspoken is that, with Ranked Choice Voting, Alameda would have seen a result like Oakland where the primary vote getter would not be the ultimate winner once the ranking reallocated lower placing candidates’ votes.

Look, I like Ranked Choice Voting, and I think that Alameda could definitely benefit from it.   I even mentioned it again when I wrote my prediction post of who would actually be running for Mayor back in February 2009.   But I think that the enactment of Ranked Choice Voting should be because we want to change the way that we vote and how those votes are calculated as opposed to being for a system because we think that it would have helped our candidate of choice that just happened to lose in the current system we have in place.  Because there are other options to either the plurality system we have in place now, or the Ranked Choice Voting system.   There is also the run-off, used at the County level for the Board of Supervisors, which is fairly simple to understand, but not nearly as powerful as RCV.

Dan Wood over at a Progressive Alamedan wrote a very interesting analysis of how Ranked Choice Voting might have affected the City Council race.    The City Council race was definitely much more complicated because it is a multi-seat election.   But let’s take the Mayor’s race for example, which is more straightforward.

I’ll caveat this by saying that I am going to make a lot of assumptions about the vote allocation, because honestly we would never know for sure how people would vote if they had the ability to cast second and third choices.   But I’m going to analyze this using a fairly black and white scenario using the premise that Doug deHaan was the spoiler in this election for Frank Matarrese.   Something that I disagree with because as I mentioned previously, this was Doug deHaan’s shot at the Mayor’s seat and not running, particularly because he clearly has an interest in the top spot, did not appear to be an option for Doug deHaan.   But admittedly, Doug deHaan voters would have been more likely to vote for Frank Matarrese than they would have voted for Marie Gilmore, however, Frank Matarrese votes would have probably been a more even break in a race of Doug deHaan versus Marie Gilmore.

Here are the final results from the Registrar of Voters:

So in a really simple analysis of the Mayor’s race, assuming that all of Kenneth Kahn’s votes would be evenly distributed amongst all other candidates because really he is just a protest vote.  In fact with RCV, I bet the total votes for a candidate like Kenneth Kahn would be even higher because of the lack of a fear of “wasting” a vote.   The next assumption is that all Tony Daysog votes would go to Marie Gilmore.  And the final assumption is that all Doug deHaan votes would go to Frank Matarrese.


Round 1 Votes Round 1 Transfer Round 2 Votes Round 2 Transfer Round 3 Votes Round 3 Transfer Round 4 Votes
Marie Gilmore
9298 +220 9518 +3197 12715 +0 12715
Frank Matarrese
6074 +220 6294 +0 6294 +6090 12384
Doug deHaan
5870 +220 6090 +0 6090 -6090
Tony Daysog
2977 +220 3197 -3197
Kenneth Kahn
882 -882

Of course, if the voting broke down in such a simplistic way, Marie Gilmore would have won in the round that Tony Daysog was eliminated because she would have been over the 50% threshold.

More likely the proportions would mirror San Leandro’s Mayor’s race, which had a 38%/32%/30% split when Sara Mestas was eliminated, and a 53%/47% split when Joyce Starosciak was eliminated.  The way I calculated these numbers was to remove the exhausted votes from the vote total and then get the percentage, in case anyone was wondering.   Although I kept the even split of Kenneth Kahn’s votes throughout all the scenarios because it’s hard to really pinpoint why voters opted for Kenneth Kahn except as a protest against all the choices.   I could have easily have eliminated him entirely and not accounted for his votes.

In round 2 where Tony Daysog was eliminated I apportioned the shares thusly: Gilmore (38%), Matarrese (32%), and deHaan (30%).   In round 3 where Doug deHaan was eliminated, I apportioned the shares as: Matarrese (53%) and Gilmore (47%).


Round 1 Votes Round 1 Transfer Round 2 Votes Round 2 Transfer Round 3 Votes Round 3 Transfer Round 4 Votes
Marie Gilmore
9298 +220 (25%) 9518 +1215 (38%) 10733 +3313 (47%) 14046
Frank Matarrese
6074 +220 (25%) 6294 +1023 (32%) 7317 +3736 (53%) 11053
Doug deHaan
5870 +220 (25%) 6090 +959 (30%) 7049 -7049
Tony Daysog
2977 +220 (25%) 3197 -3197
Kenneth Kahn
882 -882

In this sceanrio, Marie Gilmore easily wins.

Although I will say that, I think 53% apportioned to Frank Matarrese would be entirely too low if Doug deHaan were eliminated.   I bet the number of Doug deHaan voters would be much higher.   So, to better reflect what I think would have been the actual split of Doug deHaan votes, I’ve selected an arbitrary percentage of 73% for Frank Matarrese and 27% for Marie Gilmore, which means I added 20 percentage points to Frank Matarrese’s transfer and reduced Marie Gilmore by 20 percentage points.   Would Doug deHaan votes have swung 73% to 27% Frank Matarrese to Marie Gilmore, it’s possible if there were a concerted campaign amongst Frank Matarrese and Doug deHaan voters to opt for Frank Matarrese as Doug deHaan’s second choice and vice versa.   Of course, if we are considering that Frank Matarrese and Doug deHaan actively campaigned to get their supporters to opt on a ranked choice preference collectively, it would have been foolish for Marie Gilmore to not woo Tony Daysog and have his supporters rank her as number two on their ballots.   But the math would be way more complicated and would have required a lot more assumptions about those proportions.

Here’s how that scenario would shake out when I increase Frank Matarrese’s totals by 20 percentage points in the final round:


Round 1 Votes Round 1 Transfer Round 2 Votes Round 2 Transfer Round 3 Votes Round 3 Transfer Round 4 Votes
Marie Gilmore
9298 +220 (25%) 9518 +1215 (38%) 10733 +1903 (27%) 12636
Frank Matarrese
6074 +220 (25%) 6294 +1023 (32%) 7317 +5146 (73%) 12463
Doug deHaan
5870 +220 (25%) 6090 +959 (30%) 7049 -7049
Tony Daysog
2977 +220 (25%) 3197 -3197
Kenneth Kahn
882 -882

In both San Leandro scenarios, Marie Gilmore comes out ahead, even when we fudge up the numbers to give Frank Matarrese a much larger percentage than reflected in the actual San Leandro race.

Now you may be asking, well, what about if we used the Oakland numbers.   The Oakland results may be more compelling because of the lead that Don Perata had that was eviscerated by the very end when a large portion of Rebecca Kaplan voters opted for Jean Quan as their second choice.

Okay, so here is the percentage breakdown when Joe Tuman was eliminated: 44%/28%/28%, I broke it down giving Marie Gilmore the 44% because more likely than not Tony Daysog voters would opt for Marie over Frank Matarrese or Doug deHaan.   Then when Rebecca Kaplan was eliminated, her votes split 75%/25%, so I gave Frank Matarrese 75% and Marie Gilmore only 25% of Doug deHaan’s votes.


Round 1 Votes Round 1 Transfer Round 2 Votes Round 2 Transfer Round 3 Votes Round 3 Transfer Round 4 Votes
Marie Gilmore
9298 +220 (25%) 9518 +1407 (44%) 10925 +1746 (25%) 12671
Frank Matarrese
6074 +220 (25%) 6294 +895 (28%) 7189 +5239 (75%) 12428
Doug deHaan
5870 +220 (25%) 6090 +895 (28%) 6985 -6985
Tony Daysog
2977 +220 (25%) 3197 -3197
Kenneth Kahn
882 -882

Even in the Oakland scenario, Marie Gilmore would come out ahead as well.

Of course anyone can throw together some percentages and get their guy or gal ahead in the rankings.  So while Frank Matarrese voters and Doug deHaan supporters can cling on to the belief that their guy might have been able to pull out a spectacular win had RCV been in place for this election.   Using the percentages from races where no one horse traded for second place votes in the election (San Leandro) and where candidates like Jean Quan and Rebecca Kaplan urged their supporters to rank the other candidate in their number two spot (Oakland), shows that Marie Gilmore’s initial numbers would be strong enough to keep her in the lead.

About these ads

65 Comments

  1. Fun with numbers! And, especially fun when my favorite candidate comes out on top.

    I’m actually more concerned with how we elected three people to the city council given we each had only two votes. Feels a little non-democratic to me.

    Comment by Tom — November 24, 2010 @ 7:27 am

  2. Interesting analyses!

    For those who feel that the Ranked Choice Voting didn’t give people their true choice in Oakland, it’s pretty easy to re-cast the Oakland race as a June Primary and a November runoff. Perata and Quan would have likely been the top two vote-getters in June and have had to face each other in November; it’s pretty clear that Jean, being the Anti-Don, would have picked up most of the Kaplan supporters, and won the race. Same result and logic, but it took only one election to figure that out.

    I agree comment #1 above. It would be nice to have a way to include the third choice of the city council (should the seat be vacated) into the election process. I have heard since my original write-up on RCV that apparently the elimination algorithm for two-choice ballots (like our City Council) is pretty complicated, but I haven’t found any specifics on the web….)

    Comment by Dan W. — November 24, 2010 @ 7:34 am

  3. To me there’s just something basically wrong with a system that has a race with an obvious winner yet the second or third place runner gets the winner’s bouquet. Whichever way you cut it the winner doesn’t get the prize and the voters who bet on the winner gets screwed.

    Comment by Jack Richard — November 24, 2010 @ 9:19 am

  4. While you’re at it, make up some numbers to suggest what voters would have done if SunCal hadn’t spent money on phone calls and mailers. It makes about as much sense as the rest of the fantasy figures presented. What I don’t understand is why Gilmore’s supporters feel the need to keep proving that she really won. Trust me, she won. I’m not happy about it but I accept it. You can’t do anything about the circumstances under which she won, which will always leave doubt in some minds. Ranked Choice may have made a difference in the winner, but the most attractive reason for changing to that system is that it apparently reduces negative campaigning which I think all sides would have been happy to have done without this time around.

    Comment by Denise Shelton — November 24, 2010 @ 9:21 am

  5. Jean Quan victory was in large part due to the “Anyone But Don” campaign that was launched by one of Kaplin’s supporters.

    Just curious what the result would be if some group had run the same with Marie.

    1) Matresse
    2) deHann
    3) Kahn

    I really think there are pitfalls in RCV. It is not like people are voting for their three favorites in that order. Given time people will figure statisical ways to skew any election.

    Comment by frank — November 24, 2010 @ 9:35 am

  6. Jack. I agree with you, the winner is the person who gets the most votes. Its just that simple.

    One thing that we can never get rid of is the “sour Grapes syndrome” Denise I think you are experiencing it at the moment. As a Gilmore supporter I don’t feel any need to justify anything. If Frank had won the Mayors seat I would have congrdulated him and totally supported him.

    Comment by John piziali — November 24, 2010 @ 9:41 am

  7. 4:

    Denise, RCV offers us a way to reduce the impact of negative campaigns a la Perata and SunCal. I think that alone is worth putting RCV near the top of our “sunshine” agenda.

    I don’t sense any need to “prove” Gilmore’s obvious and clear win here: Lauren is just using the numbers available to her from the most recent election
    and trying to compare some hypothetical outcomes. Gilmore certainly won by a handy enough plurality on her own…

    I honestly do not believe that SunCal’s expensive campaigning actually swayed voters one way or the other in Alameda. We all reacted negatively to it and then made up our own minds, as far as I can tell. (Do you actually know of anyone who relied on SunCal’s recommendations when they voted? I sure don’t.)

    Comment by Jon Spangler — November 24, 2010 @ 10:44 am

  8. Re 7:

    (Do you actually know of anyone who relied on SunCal’s recommendations when they voted? I sure don’t.)

    Jon, do you ever get tired of making yourself sound stupid? Ever? Reeeeeeeealy? Really? Because in a room of 100 Alamedans, Spangler, you’re loud, stupid, pathetic and always wrong one.

    A friendly hint: at this point, not saying anything won’t repair the damage you have done to yourself and your reputation in this community, and neither will having sweet, sweet Linda say something venomous.

    Just go work on your tofurkey for Thanksgiving and ask the wifey to sign you up for those correspondence courses to embiggen your brain, unless you can find something other than a crayon to apply with.

    Comment by Adam Gillitt — November 24, 2010 @ 10:58 am

  9. Jon, you honestly do not believe that SunCal’s expensive campaigning actually swayed voters? Why? The only way you could prove that is if the candidates they supported didn’t win. Oh, but they did. You are kidding yourself if you think even half the people at the polls had a clear picture of how SunCal figured in any of this. Most Alamedans don’t read the blogs, the Sun, or the Journal. They don’t have friends on the City Council, or the Planning Board, or the Sunshine task force, and a lot of them would be hard pressed to find City Hall on the map. They’re not stupid, they just don’t care as much as we do. I can’t prove it, it’s just the sense I got from talking to people during the election who were not well informed.

    Comment by Denise Shelton — November 24, 2010 @ 11:06 am

  10. I’m not convinced that RCV reduces negative, venomous campaigning. In Oakland, I must have received 14 anti-Jean Quan mailers from “IEs” influenced in whole or part by Don Perata. I also received about 12 mailers that were anti-Perata from Jean Quan’s campaign directly. Alameda gave us some of the state’s most notable politicians — Don Perata, Wilma Chan, Bob Crown, George P. Miller — I wouldn’t expect your island city to be a stranger to negative campaigning or “outside influences.”

    Comment by Oakland Perspective — November 24, 2010 @ 11:36 am

  11. Jon, if “negative” campaigning is sooo hated by all the voters, why is it effective? “Well”, you might say, it’s not effective.” “Well” I might say, if it’s not effective, what difference does it make?”

    Besides that, what you call “negative” campaigning many others might call, “getting the truth out”. Furthermore if your “sunshine” committee’s agenda includes curtailment of free speech, that ain’t a sunshine committee that’s a Hugo committee.

    Comment by Jack Richard — November 24, 2010 @ 11:46 am

  12. Lauren, you speculate well. But there are no facts in your assumptions In this article. Who knows how the ranked choices really would have gone? I saw plenty of houses with deHaan and Bonta signs, so you can’t assume that people would choose Tony second to Marie, or that everyone who voted for Frank would have chosen Doug instead.

    What you do point out in this article is that it would be nice if our City had leaders that were actually supported by anything close to a majority of the population, instead of small, vocal, and fairy exclusionary, angry little factions. Bravo.

    I seem to recall someone saying something about the voice of the Citizens months ago, but I think that disappeared under a pile of money and negative campaigning.

    Comment by Adam Gillitt — November 24, 2010 @ 11:46 am

  13. I moderated a candidate’s night for the Oakland League – District 4. There were 7 candidates. All were articulate; answered the questions directly; never slammed the others on the podium. I marvelled at the contrast with our forums where some spent more time “dissing” their opponents than talking about their own positive qualities. I was told that RCV has that effect – candidates do not want to appear to be nasty or negative because that won’t get them the vote for 2nd or 3rd choice.
    I don’t think SunCal’s pathetic missives had a hill of beans effect on the electorate. Since some candidates were being slammed by other candidates and others being slammed by SunCal, everybody had an equal opportunity slamming. Most of the slamming was not very credible, no matter the source. The whole exercise was inexcusably negative, unwholesome and gave the voter nothing he or she could use to determine who had the best qualities for leadership. I think we can do far better as a community, and hope that we will.

    Comment by Kate Quick — November 24, 2010 @ 3:07 pm

  14. Are you preparing for a life in standup comedy, Kate? Or did you lose all touch with irony post-menopause? Between you and Spangler, y’all could start a commenter’s Special Olympics. Give us a break. Who are you going to be holier than next?

    Comment by Adam Gillitt — November 24, 2010 @ 3:29 pm

  15. As an outside observer and an expert on RCV (I prefer to call it instant runoff voting, or IRV), I agree with Lauren Do that one shouldn’t pick an electoral system just because you think it’ll help your favorite candidates win. What IRV/RCV will do for Alameda is ensure your Mayor, Auditor, and Treasurer really do have a majority mandate, whether or not they lead in first choice rankings. It will eliminate the spoiler effect, reduce “lesser-of-two-evils” voting, and minimize the impact of vote-splitting, where the majority’s vote is split amongst too many similar candidates.

    If applied to the City Council seats, instead of the majority of the voters taking both seats each time, the multi-winner cousin of IRV, Choice Voting, would ensure fair representation for any electoral minority that constitutes at least one-third of the voters (assuming they all vote cohesively). (If that’s too much democracy for some of you, note that one could enact IRV for the single-winner races and leave the City Council races as they are.)

    Concerning #10′s comment on negative campaigning, no electoral system will eliminate it entirely, but IRV does reduce it. For example, in the Oakland Mayoral contest, Quan and Kaplan cross-endorsed each other, each asking their supporters to rank the other one second. In the first election of a traditional two-round runoff system (as Oakland used to use), they would have not done that, and probably would have attacked each other as well as the other front-runners (Perata and Tuman) in a desperate attempt to be amongst the top two vote-getters.

    Members of my organization, Californians for Electoral Reform ( http://www.cfer.org), were part of the coalitions that brought IRV to Oakland, Berkeley, San Leandro, and (in 2004) San Francisco, and we’ll be happy to help any Alamedans(?) who want to bring it to their city. (While we do have an East Bay chapter, we currently have no members who live in Alameda itself. We can advise, but we cannot conduct a campaign for you.) I personally gave a presentation on IRV to the Alameda League of Women Voters in 2006, and would be happy to reprise that for any group.

    Comment by Steve Chessin — November 24, 2010 @ 3:36 pm

  16. Adam. I thought by now you would be busy moderating your blog. after all you have had one comment since it started. I guess you have to post here, as the policy on your blog would not allow you to post there.

    Comment by John piziali — November 24, 2010 @ 4:00 pm

  17. If you want to change the voting process in non-partisan elections, why not change it to range or score voting? Each voter can score a candidate, let’s say between 1 and 10 on the ballot, like competitors in gymnastics. Hard to tell who the best gymnast is but somewhat easier to spot the range of the best. The candidate with the highest average would win.

    Comment by Jack Richard — November 24, 2010 @ 5:47 pm

  18. Jack (#17): Range voting allows a minority to out-vote a majority, even in a two-person race. Consider a race between Jill and Joe, with 10 voters. Two (Jill and her significant other) really like Jill and can’t stand Joe, so score her 10 and him 1. The other eight voters don’t like either of them much, but have a slight preference for Joe, so give him 2 and her 1. (Joe forgets to vote.) She gets 28 points (average of 2.8), he gets 18 (average of 1.8), so she wins, even though a majority preferred Joe to Jill.

    Granted, this is an extreme example, but one can construct less extreme examples where the majority preference loses.

    For more on range voting (and other systems) as compared to IRV, see http://www.fairvote.org/how-instant-runoff-voting-compares-to-alternative-reforms/

    Another critique of range voting can be found at
    http://archive.fairvote.org/rangevoting.pdf

    Comment by Steve Chessin — November 24, 2010 @ 6:36 pm

  19. Too bad the Olympics haven’t discovered how unfair range voting is.

    Would it work in non-partisan multi-winning candidates, like for instance, council elections.

    Comment by Jack Richard — November 24, 2010 @ 7:03 pm

  20. Wait a minute. After thinking about your #20, I’m not buying your reasoning. If there were only ten voters, yes the election could be skewed by a few gamers. But with 20,000 + wouldn’t the gamers on all sides cancel themselves out?

    Comment by Jack Richard — November 24, 2010 @ 7:14 pm

  21. Steve, my 20 was aimed at you not me.

    Comment by Jack Richard — November 24, 2010 @ 7:23 pm

  22. Re: #19. Actually, they have. See, for example,
    http://en.wikipedia.org/wiki/2002_Olympic_Winter_Games_figure_skating_scandal

    So now they use a different system, where they have 12 judges, they randomly pick nine, they throw out the highest and lowest value, and average the rest. See
    http://en.wikipedia.org/wiki/ISU_Judging_System

    Works for skating, I suppose, but I wouldn’t recommend it for a public election.

    Comment by Steve Chessin — November 24, 2010 @ 7:28 pm

  23. Re: #20. (I know you meant my #18.) That assumes there are the same number of gamers on both sides.

    So let’s take a less extreme example. Of your 20,000 voters, 45% (9,000) strongly prefer A to B, so score A 10 and B 1. The other 55% (11,000) prefer B to A, but less strongly, so they score B 8 and A 3. A gets 123,000 points, B gets 97,000 points. A wins, even though a clear majority preferred B.

    Besides, don’t you want a system that encourages voters to vote sincerely, so they don’t have to game the system?

    I strongly urge you to read the paper at
    http://archive.fairvote.org/rangevoting.pdf

    Comment by Steve Chessin — November 24, 2010 @ 7:39 pm

  24. 9. 11:

    Denise, I DID talk to lots of voters during the campaign – and most of them (at the Farmer’s Market, Webster Jam, and in canvassing) were, indeed, very well informed. They knew about SunCal’s mailers and made their choices independent of them, or, in a few cases, had decided to vote for the candidates that SunCal had targeted.

    In the latter case I had my work cut out for me to defend the honorable people who had been smeared by SunCal’s “endorsement” (Gilmore, Tam, Bonta).

    It’s easy to confuse the several independent variable involved in this elections: 1) SunCal’s expenditures and campaigning, 2) the effects (if any) of those efforts, 3) the actual qualifications and stands of the candidates, and 4) the voters’ ability to discern the differences between the above.

    I think you are not giving the voters of Alameda – or CA – sufficient credit for their intelligence and ability to tell the difference between wheat and chaff. After all, we said no to Whitman and Fiorina and Proposition 23….

    Comment by Jon Spangler — November 24, 2010 @ 8:14 pm

  25. We had heard in the Democratic Party circles that there was a young man that attacked an elderly woman in a wheelchair at the campaign headquarters. Do you think RCV or IRV would have changed his behavior?

    Comment by Oakland Perspective — November 24, 2010 @ 8:28 pm

  26. Funny, I heard something similar months ago… Hasn’t anything happened since then? Now pass the yams or I’ll break both your arms, and I can’t fergit my speshul Thanksgiving shoutout to my man-crush John P.!

    Comment by Adam Gillitt — November 24, 2010 @ 8:52 pm

  27. I hope everyone has a great Thanksgiving holiday,
    whatever your persuasion – light meat, dark meat, cranberry relish or cranberry sauce, sweet potatoes or green beans, football or no….

    Comment by Jon Spangler — November 24, 2010 @ 9:05 pm

  28. post 25. Oakland Perspective. nothing has changed his behavior yet. He is defined by his actions.

    Comment by John piziali — November 24, 2010 @ 9:11 pm

  29. 23
    I did read your links The argument against using the score ranking method in partisan races, where two parties dominate the spectrum, may not be appropriate as I stated above and you inferred in your #23 (which apparently assumes only two parties since 10 to 1 yay or nay would most likely only happen in a democrat/republican election, in that case the results would end up no different than a standard presidential election that we currently enjoy). But
    what about non-partisan multi-seat elections, like our
    recent council election? I doubt most voters had very strong feelings about any single candidate and if they did they were probably nullified by the same few who had strong feelings about another.

    If the purpose of the method of voting is to keep strong feeling voters from voting, I think RCV might do the trick since, all kinds of Machiavellian machinations by candidates jockeying for position instead of trying to win takes place and voters trying to figure out their third place candidate is more than they want to contend with.

    The figure skating example, again, has few voters

    Comment by Jack Richard — November 24, 2010 @ 10:11 pm

  30. &#$%*(!Russian judges!

    Comment by Denise Shelton — November 24, 2010 @ 11:43 pm

  31. No perfect voting system exists; the result is from economics, but it applies to any voting situation. Kenneth Arrow won a Nobel Prize for it (the “fake” economics one, but still pretty good). “A Difficulty in the Concept of Social Welfare” or Arrow’s Impossibility Theorem.

    Comment by elliott.gorelick — November 25, 2010 @ 8:03 am

  32. 31
    No perfect voting system may exist but one man one vote has seen this nation through the years pretty well. So why, all of a sudden does it have to be changed to one man one vote to, however many candidates are running, votes? I’d give odds that if the Oakland election was rerun now, Perata would win going away.

    Perata was favored to win and the voters who voted for him expected him to win, thus they didn’t particularly care who came in second. Most likely, since we’re talking about liberals here, Perata’s supporters honestly chose a person who they thought was best suited for second place and accordingly voted for Quan. Quan’s supporters, on the other hand, knew Perata was the leading contender and the best and only chance their candidate had was to vote Perata in last place.

    Perata’s voter’s got juked by the loser. In future elections of this sort, expect the leading contender’s voters to vote for the second place contender in last place, regardless of their true feelings. With both the first and second candidate skewing their second choice to last place the winner will be chosen by the other voting losers. And thus we march in to the brave new world where losers are the real winners.

    Comment by Jack Richard — November 25, 2010 @ 8:50 am

  33. 31
    Setting your Nobel Prize winning by Arrow’s Impossibility Theorem aside, (though, he may agree that the Nobel committee collective eshews his own reasoning) couldn’t you have used better and more recent examples of Nobel prize folly? Especially in the arena of Peace?

    Comment by Jack Richard — November 25, 2010 @ 9:20 am

  34. Jack, maybe I don’t understand RCV but I think in the end one man or woman still gets just one vote, though in RCV it counts where as in plurality votes can be wasted or a vote for Nader by a person preferring Gore over Bush can in effect get Bush elected.

    Perata’s supporters didn’t have their second ballot counted so if they put Quan as second choice it didn’t matter. When you cast a vote for a candidate who is deemed incapable of winning, your next choice is tallied. Voters whose first ballot was for a person in the run-off have their vote stay with that candidate. If after round two the second choices are still not for one of the two front runners, then your third choice comes into play, and so on until somebody gets over 50%. In some RCV ballots you can rank all candidates, even twelve if there are that many. Usually 3 is enough to get the same result and voters new to the system just couldn’t cope with a fully ranked and comprehensive RCV ballot.

    Comment by M.I. — November 25, 2010 @ 10:42 am

  35. Wow, it’s even more complicated than I thought. So the winning candidate’s supporters’ extra votes don’t count. Only the losers votes count choosing the winner. This is rank.

    Comment by Jack Richard — November 25, 2010 @ 11:02 am

  36. I guess the theory is, since the losers wasted their vote on losers in the first round and in a separate runnoff election the losers would still select the winner since the 1st and 2nd place voters probably wouldn’t change their votes so, save the money and count the losers twice in the first round.

    Comment by Jack Richard — November 25, 2010 @ 11:16 am

  37. Jack,,

    I think you are being unnecessarily contrary. There is the added layer of campaigning where candidates can do “coalition building”, but really this is just a simple math equation which save the cost of a run-off. Even I can see that.

    If Perata did win in a run-off it would be because it was a different election in which he could start from scratch and use his greater war chest to beat crap out of Quan, where as on Nov 2 Quan was the real choice of majority of people, had just she and Perata been in the election that day.

    The simplest way to look at it is that after nobody got a majority, and Kaplan as 3rd place was eliminated from any run-off election scenario, then the run-off data had already been collected and the Kaplan people simply voted for Quan.

    I don’t know how people can get so hung up on the fact that because Perata got a plurality in the first round that somehow means nobody else could get the ultimate majority. Quan did, and it’s completely fair and logical.

    Comment by M.I. — November 25, 2010 @ 1:00 pm

  38. Mark, all my contrariness is necessary.

    Part of that contrariness is my questioning of why the candidate who eventually holds the seat must have, somewhere along the line, received more than 50% of the voters votes. Perata did get more of the votes than anybody else but he did not hit the 50% + one of the vote’s magical marker number. So, what RCV does is contrive a way to wave the clear winner’s honest win in favor of a 50% plus one majority by allowing some voters two votes.

    If there was a runoff election between the two highest vote getters (which I say is unnecessary, since someone already won the race), darned right it should be a new election where either candidate could use whatever resources available. The cost of the runoff election is the unnecessary cost of a 50% plus one requirement when someone already won.

    Comment by Jack Richard — November 25, 2010 @ 3:03 pm

  39. #38: Jack, I don’t know if you live in Oakland or not, but even I (a non-resident) know that Oakland used to have a two-round runoff system, with the first election in June. If someone got over 50% in the first election, they were elected, end of story. If no one got a over 50% in the primary, the top two vote-getters went to a runoff in November. Sometimes the person who was first in the primary was also first in November, but sometimes not. (For example, when Jean Quan ran for the District 4 City Council seat in 2002, she finished second behind David Stein in the March primary but beat him in the November runoff.)

    With IRV (RCV) it’s the same thing. Sometimes the person who leads in the first round wins (for example, City Council District 4), sometimes not (Mayors race).

    You may question the need for runoff elections (instant or delayed), but the City Charter requires it.

    Also, no one gets more than one vote. Everyone one gets one vote, it’s just that some votes get transferred from one candidate to another, just as, in two-round runoff, if the candidate you voted for in the first election isn’t in the runoff, you get to pick someone else. With RCV you just pick your second (and third) choice before the first (and only) election, instead of after.

    Comment by Steve Chessin — November 25, 2010 @ 10:08 pm

  40. #29: Jack, I used a two-candidate example not to illustrate partisan races, but to show that, even in a two-person race, score voting violates the simple principle of majority rule. Clearly, in a two-person race, RCV does not.

    In multi-candidate non-partisan races, with three or more viable front-runners, once voters realize that giving any score below the maximum to your first choice will hurt your first choice (as will giving any score above minimal to the other candidates), they will give 10 to their first choice and 1 to the others. This will result in plurality winners.

    For example, consider a city council race with three candidates, Jill, Joe, and John. Forty percent strongly prefer Jill as first choice, thirty-five percent strongly prefer Joe, and twenty-five percent strongly prefer John. Using score voting, with each voter giving 10 points to their preferred candidate and 1 to the others, Jill wins, even though she is preferred by only a plurality of the voters. She *might* be preferred by a majority, or Joe might be, but score voting doesn’t let us discern this.

    With RCV, because ranking someone second does not hurt your first choice, voters have no disincentive to vote sincerely, and the second choices of the voters for John will let us know if Jill or Joe has the broader support.

    There’s a ranked system called Bucklin that Alabama briefly used. In Bucklin, if no one gets a majority of first choices, all the first and second choices were counted, and whoever has the most votes wins. Voters quickly realized that this meant your second choice could hurt your first choice, so stopped ranking second choices, and Bucklin degenerated into simple plurality. (See http://archive.fairvote.org/?page=2077 for more information on this.)

    But there are other arguments against score voting made by the paper I referenced in #23. While the scale seems linear, it really isn’t; the psychological distance between 4 and a 6 is a lot smaller than that between an 8 and a 10. And different individuals, even ones with identical politics and values, will score candidates differently. While each may agree that the one they like best should get a 10 and the one they like least should get a 1, assigning scores to the ones in between is more complicated. A candidate scored 6 by one might be scored 4 by another. (See the examples in the cited paper.) Combining non-comparable numbers gives a meaningless result.

    Comment by Steve Chessin — November 25, 2010 @ 10:44 pm

  41. sorry folks, I still don’t see what’s wrong with our election process here in Alameda. We have people run for office and we vote for those who run on a given day, the person who gets the most votes on that day wins. Its kind of like a foot race fastest person wins.
    Maybe our parcel tax election should be set up the same way, yes or no, most votes win. simple.

    Comment by John piziali — November 25, 2010 @ 10:47 pm

  42. #41: John, in most “yes or no” elections (excluding parcel taxes and other measures that require a 55% or greater super-majority), just as in a single-seat contest with only two candidates, “most votes win” is identical to majority rule.

    But in a single-seat contest with three or more candidates, “most votes wins” can result in someone winning who is *opposed* by the majority of the voters.

    Consider your Mayor’s race. Marie Gilmore won, fair and square, with only 36.9% of the vote, because your charter only calls for “most votes wins”. But this means her political opponents can say, loudly and often for the next four years, “63.1% of the City voted against her; she doesn’t have a mandate”, which can hurt her effectiveness as Mayor. (A response of “but she got the most votes” is weak, at best.)

    Now, according to the analysis of the results by Lauren Do at the top of this blog post, and the assumptions she makes, a majority of the voters probably do prefer her to the other candidates, but we’ll never know for sure because we have no true information concerning the second and third choices of the voters, only Lauren Do’s conjectures.

    Had Alameda used RCV in the Mayor’s race, Marie Gilmore and the voters would know if she had a mandate, and how large that mandate was. While her political opponents would still be able to say “63.1% of the voters would have preferred someone else as Mayor”, she’d be able to respond (assuming that she would still be the winner with RCV) “but I was able to forge a majority coalition of the voters and none of the other candidates were able to do that.” That’s a much stronger response.

    (And if someone else would have been the winner had this election been conducted using RCV? Well, if you believe in majority rule, wouldn’t that be the best outcome for Alameda?)

    Comment by Steve Chessin — November 25, 2010 @ 11:51 pm

  43. 38., 41. I think Steve is being crystal clear about how under RCV voters only get one vote, and how the run-off is instant instead of protracted and expensive. The previous election was exhausting and I’m glad it’s over, but I still prefer most democratic voting method which is majority rule, and the most efficient method for that, which is RCV.

    I don’t know how you guys can still stick by plurality. If we had enough candidates we could elect a mayor with 12%. That’s messed up.

    #5. there are limits to how far the election can be skewed and I don’t think they are as bad as a candidate with a huge war chest jockeying to get into a run-off so they can smother the run-off opponent.

    Comment by M.I. — November 26, 2010 @ 9:24 am

  44. Thanks Steve for your comments concerning my, admittedly, unlearned analysis of alternative voting processes. I do live in Alameda, have for 38 years and am not advocating for any change in Alameda’s method of voting system. Like John, I don’t see what’s wrong with the current election process. Perhaps that opinion might change if Alameda had a District instead of an At Large system.

    I would make a couple observations on your response to John, in #40. As I understand it, your support of RCV, at least in your hypothetical example of Gilmore getting less than 50% of the total vote, stems from the definition of the word “majority”. It’s your contention that unless a winning candidate gets more than half the total votes, the elected official’s effectiveness will be hurt.

    I believe that the effectiveness of a candidate, regardless of whether that candidate gets 49 or 51 percent of the vote, will be determined and judged by the candidate’s performance in office. The total numbers of electors, who voted for a candidate, are quickly forgotten by citizens. Sure there are those political opponents who will attempt to capitalize on any and all reasons to try to denigrate their opponents but, by and large, most voters, I believe, are sophisticated enough to discount those type criticisms.

    So what it boils down to, again, at least in your argument, is the definition of the word “majority”, as opposed to the word, “plurality”. In your # 39, you state, “…no one gets more than one vote. Everyone one gets one vote, it’s just that some votes get transferred from one candidate to another,…”. Okay, I’ll accept that everybody only gets to vote once, but contained within that one vote, at least for the biggest loser voters, is the potential to have two votes count. The first vote that joined those few other last place candidate votes and the second vote that will potentially wrest the victory from the candidate who finished the race with the most votes and give it to the next in line.

    The word “majority” is muddled in many English speaking countries, from what I’ve found. From the definition in:
    Cambridge Advanced Learner’s Dictionary

    majority noun ( NUMBER )
    /məˈdʒɒr.ə.ti//-ˈdʒɑː.rə.t ̬i/ n

    [S] the larger number or part of something

    Comment by Jack Richard — November 26, 2010 @ 9:31 am

  45. Sorry Mark and Steve, I’m still with Jack.
    I will always vote no matter what system we use. My problem with this election is some people want to change the rules because they are bitter in their loss. Another words they are not looking for change in a positve way but somehow hoping it could change the outcome of an election to their benifit.
    Being Mayor of the city of Alameda should not require a mandate as the seat gives the holder no extra power. Our mayor simply runs the meetings and cuts ribbons. Each council member has but one vote. As a matter of fact the Mayor can’t even appoint someone to a board or commission without approval of the rest of the council.

    Steve, in reference to your last sentence in post #42. yes if the election would have been conducted in that manner I would have accepted the vote and moved foward, not tried to have the rules changed because I was unhappy with its outcome.

    Comment by John piziali — November 26, 2010 @ 9:55 am

  46. 45. I think you’re wrong about people wanting to give the new system a try because they’re bitter. Nothing will change or adjust the results of the recent election and a new system won’t favor one ideology over another, so it would be silly to think “if only we had RCV, my candidates would win.” The reason I’m interested in it is because this election was so contentious and the RCV system promises to be less so. It would also speak to those who feel the winner of the mayor’s race should have a higher percentage of the voters’ endorsement. If the effort to try RCV has been inspired by sour grapes, so what? That doesn’t reflect negatively on its value as a system. Changes always occur when someone is unhappy about something. That doesn’t mean the rest of us won’t benefit.

    Comment by Denise Shelton — November 26, 2010 @ 10:22 am

  47. So, Denise, you’re bitter with the “contentiousness” of the recent election and want to change the process because somehow, somewhere along the line, someone declared that the particular system you want to change to will eliminate “negative campaigning” (whatever your definition of that is), right?

    What happened in the Oakland election is, the winner of the race was replaced with a non-winner because the biggest voting losers had their votes counted twice. If I were a Perata supporter in Oakland, I would be far more bitter with the outcome of a process that can make my winning candidate a loser by voters who’s second choice was a loser and who’s first choice was the biggest loser.

    And, “giving the new system a try” means living with it, good or bad, for a long, long time.

    Comment by Jack Richard — November 26, 2010 @ 11:48 am

  48. 44. I noticed I subtracted ten years off my life in Alameda. I should’ve written 49 years instead of 39. I guess it’s just that living here makes you want to stay longer. So I dedicate this to others entering senectitud.

    Comment by Jack Richard — November 26, 2010 @ 12:21 pm

  49. Denise I’m not just singling you out. But lets be honest here, if one of the other candidates had won the Mayors seat do you still believe we would be arguing these points.

    Just a question, why are we only arguing about the Mayors seat and not the other two council seats in question here. Like I said just a question.

    Comment by John piziali — November 26, 2010 @ 12:22 pm

  50. Will you all stop characterizing me as bitter? I am not bitter. I don’t have the energy for bitterness. I’m saving bitter for later in case I need it.

    Comment by Denise Shelton — November 26, 2010 @ 1:29 pm

  51. #43: MI: Yes, right, exactly.
    #46: Denise: Yes, right.

    Perata did not get “the most votes” — he got a plurality of the first choice votes. Of all the votes cast for the two top vote getters, Quan got the majority (which in voting means a threshold of 50%, which is why “super majority” means a threshold of greater than 50%, right?)

    Every time a first choice vote for a losing candidate is eliminated, that vote is negated and then the voter’s second choice vote comes into play. It’s still one vote.

    I’m going to guess that people opt for RCV more often than not for the wrong reasons, but so what? It’s still an inherently fair system and it’s obvious to anybody w/ common sense that it doesn’t guarantee any particular result.

    I would only note that the city doesn’t require a majority (50%+) vote — it’s only a plurality and I think that’s in the charter. So assuming that RCV only makes sense with a winning threshold that’s at or near a majority, then I guess an RCV system would requirement a charter amendment, for the mayor’s race anyway.

    Comment by dlm — November 26, 2010 @ 2:24 pm

  52. 44. Jack,

    Can you possibly be serious about performance being the test of an election? Performance is entirely separate from mechanics of an election. How well an elected candidate performs has nothing to do with the method of election. Should Oakland argue against it’s pre-IRV system because Ron Dellums was invisible for four years?

    This island is too small for district elections. We could have north, south, east, west, but there would likely be a shortage of qualified candidates if each had to live in a given district.

    As noted before the RCV method for multi-seat races, Single Transferable Vote, has been shown to mathematically render an effect similar to districts in that it tends to give more representation to constituents which are underrepresented in our current system. I’d need to dig up a 13 minute power point I had which explains this.

    49. John,

    I believe Denise about outcome. I’ve been behind RCV for several years despite prospects for my preferred outcome, but it RCV was a total non-starter until we had a contentious campaign where there were guaranteed to be dissatisfied voters one way or the other. I’m put off by editorials and letters about how we should do this because somebody didn’t like the outcome, but what else ever motivates people to act?

    I don’t know that it’s needed, but I’m bitter about almost everything at one time or another and it definitely takes it’s toll.

    Comment by M.I. — November 26, 2010 @ 2:36 pm

  53. 52 Mark, read this carefully, “I believe that the effectiveness of a candidate, regardless of whether that candidate gets 49 or 51 percent of the vote, will be determined and judged by the candidate’s performance in office.” Jack Richard #44

    Now carefully read your question directed at my statement, “Can you possibly be serious about performance being the test of an election?”

    Now, do you know what the difference is between a, “candidate” and the, “election”?

    And, I never said I was for district elections in Alameda. I agree with your comment about “district” elections so please don’t imply that I’m for something I never said I was for.

    I don’t mind discussing issues with you but please don’t add, subtract or try to change the meaning of my words without my prior approval.

    Comment by Jack Richard — November 26, 2010 @ 3:55 pm

  54. 50. You didn’t say who you were directing your admonishment to in #50, Denise. In case it was my #47, I must say it wasn’t my intent to mischaracterize you. I inferred that the negative campaigning (your definition) you spoke of several times had left a sour taste in your mouth. If I was wrong, please accept my apologies.

    Comment by Jack Richard — November 26, 2010 @ 5:23 pm

  55. yeah Jack. good point, sloppy reading. I don’t need to be scolded.

    However did you not make a reference to being open to district elections if we didn’t vote by plurality? (44 first paragraph) If not, what did you mean?

    If you want to exchange admonishments for putting words in each others mouths, I never said you said you said were for district elections.

    Comment by M.I. — November 27, 2010 @ 4:27 pm

  56. Yes, I am open to all things but all that means is my mind isn’t closed.

    Tell you what Mark, I watched this clip and it really made me realize how easy it is to assign guilt when none may be there. I’d recommend it to anybody.

    http://www.wimp.com/judgejudy/

    Comment by Jack Richard — November 27, 2010 @ 4:49 pm

  57. I don’t know, Jack. I think she broke the toilet. I don’t think she should pay for it, though. This actually happened to me when I had a plus-sized friend use my bathroom in London. But I didn’t make her pay for it. If figured the humiliation was bad enough.

    Comment by Denise Shelton — November 27, 2010 @ 8:18 pm

  58. This conversation is missing something. If Frank or Doug had won would anyone be talking about ranked voting? answer: I think not.

    Comment by Hot R — November 28, 2010 @ 8:48 am

  59. Jack, what really made watching this hysterical is that on my computer the tape speed was slow so the effect was as if everybody were very drunk. It was so slow I only got to “She musta’ fell on it.”

    58. I may be a broken record here, but what’s new? A small group of people I know have been beating this drum for some time and will continue. Look to LWV for some leadership in the future. In fact when the county was shopping for new voting machines Joanne McKray from LWV invited me to help lobby Alice Lai Bitker to get the County Registrar to buy machines with RCV capacity. The irony is that even though we succeeded on RCV, because of the whole electronic voting machine scandal, those machines were scraped.

    I don’t deny that the reason RCV is getting play now is the outcome of this race, but that doesn’t change the basic validity of the system. I’m glad the shoe is not on the other foot because if my candidates lost I would be consistent in advocating for RCV on principle, but nobody would believe that.

    Comment by M.I. — November 28, 2010 @ 9:32 am

  60. A agree with you Denise, I think the plaintiff’s plus-sized friend probably leaned back against the tank. Leaning on the tank put excessive pressure on the bolts that fasten the tank to the bowl and that cracked and broke the porcelain tank and possibly the bowl too. But the plaintiff brought no proof that happened so judge Judy, in all her slow ponderous explanation, made the right decision.

    59. Mark, maybe it’s the operator’s system that’s slow.

    Comment by Jack Richard — November 28, 2010 @ 10:24 am

  61. never as quick as you Jack, of course.

    Comment by M.I. — November 28, 2010 @ 10:49 am

  62. http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2010/11/27/INQT1GEP3A.DTL

    part of my response:

    His use of the San Francisco District 10 race as an example of IRV failure was disingenuous. He decries the fact that the person with most first round votes lost to a candidate who got a mere 11.8% in that round but eventually prevailed. In fact that 11.8%, cast for Tom Kelly, was the highest percentage for any candidate in round one. The eventual run-off victor Malia Cohen received 11.78% in round one. The reason that no candidates started off with a stronger showing was that there were 21 candidates in the race!

    In a traditional run-off, two candidates with less than 12% of the vote would have gone head to head, immediately eliminating the rest of the field. In IRV the votes of last place finishers of each round are redistributed one round at a time, allowing the choice of a true majority of voters to gradually emerge through a methodical elimination of lesser candidates one by one.

    Comment by M.I. — November 28, 2010 @ 6:11 pm

  63. 62.
    “In IRV the votes of last place finishers of each round are redistributed one round at a time, allowing the choice of a true majority of voters..”

    True majority my ass. The true majority were those voters who voted for the candidate who got the most votes in the first and (should of been) only round.

    What’s next, jimmying the results of a hundred yard dash so somehow the last place finisher decides who won the race?

    Comment by Jack Richard — November 29, 2010 @ 8:37 am

  64. No, it’s letting the race’s sponsor decide the finishing order, Jack, come on. Dont be so naive.

    Comment by Adam Gillitt — November 29, 2010 @ 10:29 am

  65. #63. Jack, while some dictionaries give “most” as one definition of majority, the one that applies in the context of electoral systems is “more than half”. The front runner in round 1 in San Francisco’s District 10 race was Lynette Sweet with 12.07% of the vote. That’s less than one-eighth of the vote, no where near a majority.

    And it’s not the “last place finisher” who decides the contest. It’s the majority of the voters.

    Comment by Steve Chessin — December 31, 2010 @ 3:55 am


RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

The Silver is the New Black Theme Blog at WordPress.com.

Follow

Get every new post delivered to your Inbox.

Join 730 other followers