What politicos can learn from college football
Kari Chisholm
Now and then here at BlueOregon, we've talked about various kinds of reforms to our voting system -- instant run-off voting, approval voting, proportional representation, etc. (Not to mention the extensive discussions about the top-two primary.)
Just this last weekend, there was a great example of a national election in which the choice of voting system defined the outcome.
It was an election between three strong candidates (and at least a dozen minor candidates). Going into the vote, no one had any idea who was going to win. And, most intriguingly, there was a strong regional effect - with two of the top candidates coming from the same region.
An election? Last weekend? Yup. The Heisman Trophy - the most prestigious individual honor in all of sport - is awarded through a vote of 870 sports journalists and 55 former winners.
Voters were asked to name their top three candidates, and rank them in order. Each #1 vote was worth three points, each #2 vote two points, and each #3 vote one point.
The three finalists were Sam Bradford of Oklahoma, Colt McCoy of Texas, and Tim Tebow of Florida. In one of the closest three-way votes in history, the winner was Oklahoma's Sam Bradford.
But here's the thing: If that vote had been held according to the rules we use for most of our elections - vote for one - the winner would have been the third-place candidate, Tim Tebow. Despite placing third, he had the most #1 votes.
Here's the vote breakdown:
| #1 | #2 | #3 | total | |
| Bradford | 300 | 315 | 196 | 1726 |
| McCoy | 266 | 288 | 230 | 1604 |
| Tebow | 309 | 207 | 234 | 1575 |
Which illustrates one of the strengths of a system that gives voters the ability to express their will in a more richly detailed way. In the Heisman vote, while Tebow got the most #1 votes, there were more people (388) who thought he was either #3 or worse. While Bradford fell barely short of beating Tebow in #1 votes, he had well over a hundred more #2 votes. In other words, it was close, but Bradford was the consensus pick.
And it doesn't have be a point-scoring system like the Heisman system to generate this kind of richer expression of the voter's will. Over at StiffArmTrophy.com (another site I run), I was able to publish the actual ballots of several hundred Heisman voters who disclosed their votes. And in an analysis of those votes, I found that if the Heisman vote were conducted like an instant-runoff-vote election, Sam Bradford also would have won.
Of the 207 complete ballots I know about, 78 selected Tebow #1, 67 for Bradford, and 62 for McCoy. Dropping McCoy from the mix and shifting his voters to their second-choices, the outcome shifts to 112 for Bradford and 95 for Tebow -- Bradford wins.
And a large part of the reason for that is the regional split between Oklahoma and Texas. Voters in that region may not have agreed on who was #1, but a large number were certain that the kid from Florida was their third choice.
And that's the point of an instant-runoff vote. It prevents a spoiler effect from happening in a multiple candidate race when the candidate mix is unbalanced - either regionally or ideologically.
In 2006, a lot of Democrats were terrified by the prospect of Ben Westlund running an independent campaign for Governor. Not because we didn't like him, or his progressive/independent message, but rather precisely because we DID like his message -- and we were worried he'd split the Democratic vote with Governor Kulongoski, electing Republican Ron Saxton.
But in an instant runoff vote, that worry ceases to exist. You can have more candidates, with a broader diversity of issues and ideologies, without worrying that two of the candidates will split the support of the majority, allowing someone to get elected who doesn't represent the most possible people.
We should seek to have balloting systems that allow each voter to express their will in as rich detail as possible, ensure that the majority opinion is expressed in the outcome, and where more candidates running is always a positive thing.
It seems to me that instant-runoff voting is one great way to get there.
 |
More Recent Posts | |
Albert Kaufman |
|
Guest Column |
|
Kari Chisholm |
|
Kari Chisholm |
Final pre-census estimate: Oregon's getting a sixth congressional seat |
Albert Kaufman |
Polluted by Money - How corporate cash corrupted one of the greenest states in America |
Guest Column |
|
Albert Kaufman |
Our Democrat Representatives in Action - What's on your wish list? |
Kari Chisholm |
|
Guest Column |
|
Kari Chisholm |
|
connect with blueoregon
2:24 p.m.
Dec 15, '08
Kari said: it was close, but Bradford was the consensus pick.
My friend, I beg to differ. I don't have a dog in this fight, but I think Tebow got robbed. If the most voters thought he was the best, I think he should have won.
"Consensus" is clearly the wrong word here. Just because a bunch of Big 12 voters got together and manipulated the system by keeping Tim Tebow off their ballots entirely (what he's not even the third best college player?) to ensure that one of the Big 12 guys won, does not make Bradford the "consensus" anything. Bradford won under the rules in play, just as Oklahoma and Florida get to go to the BCS title game under a complicated formula, despite the fact that USC is arguably the best team in the country.
As long as we are going to live in a system where we elect people to single member districts (one governor, one state senator, etc.), for my money I think the person that the most people like best should win. Tebow was robbed!
3:13 p.m.
Dec 15, '08
If the most voters thought he was the best, I think he should have won.
I agree. But he didn't. More voters thought someone else was better.
I'm not particularly interested in debating the merits of the Heisman candidates here -- well over a thousand comments on that subject at my other site -- but I do think this example helps us debate the question: what does it mean when we say "the most people like best should win"? What does "most" mean and what does "like best" mean?
As you note:
As long as we are going to live in a system where we elect people to single member districts (one governor, one state senator, etc.), for my money I think the person that the most people like best should win.
That's true -- but why should the decision-making process completely ignore the rich detail contained in a voter's will?
Let's try another example: 100 of us are trying to decide which kind of ice cream to have at a party. 70 of us absolutely love chocolate - and 30 are committed to strawberry. The choices available are: chocolate mint chip, chocolate & peanut butter, plain chocolate, and strawberry.
30 prefer #1 strawberry, #2 plain chocolate, #3 chocolate mint chip, #4 chocolate & peanut butter.
25 prefer #1 plain chocolate, #2 chocolate & peanut butter, #3 chocolate mint chip, #4 strawberry.
23 prefer #1 chocolate mint chip, #2 plain chocolate, #3 chocolate & peanut butter, #4 strawberry.
22 prefer #1 chocolate & peanut butter, #2 plain chocolate, #3 chocolate mint chip, #4 strawberry.
The greatest number of voters will have the greatest satisfaction if the choice is plain chocolate. But under a one-vote-only rule, then the winner will be strawberry.
In an instant runoff-vote: After the first round, choc & pb is eliminated. Their vote moves to their second choice, plain chocolate. In the next round, choc mint chip is eliminated, and those votes move to plain chocolate. Ta-da! Plain chocolate wins.
This seems self-evident to me: Why should strawberry be the winner when it's the dead-last choice of 70% of the electorate?
Dec 15, '08
Instant Runoff Voting is certainly not a way for voters to express their preferences in "rich detail". It ignores a huge amount of information, and gives voters an incentive to betray their favorite candidate the same way a lot of Nader supporters actually voted for Gore. When you factor in the complexity of IRV, the situation is even worse.
Score voting an approval voting are superior in virtually every way, and are much simpler.
IRV is essentially the worst of the five commonly proposed alternative voting methods.
3:33 p.m.
Dec 15, '08
Clay -- I like range-voting, too. My point wasn't to make an unequivocal endorsement of IRV as the best alternative among the choices -- but rather to argue that we ought to be considering alternatives to the current system.
(You're deep in the reform world, arguing among alternatives, while our more general-interest audience here hasn't yet decided that ANY alternative is better than what we've got now.)
I talked about IRV in my post because I had actual ranked-choice ballot data from Heisman voters. I don't have any idea what the numbers would be if they were asked to provide range-voting scores.
<hr/>For those who have no idea what Clay is talking about, a range voting ballot would work like this:
"On a scale of 0-10, score your approval of each candidate - where 10 is most-favorable. You may give the same score to multiple candidates, if you choose."
The candidate with the highest score wins.
5:59 p.m.
Dec 15, '08
Kari, excellent post. Will you join with the Pacific Green Party and Independent Party to lobby the Legislature to adopt IRV for Oregon elections?
Dec 15, '08
... a bunch of Big 12 voters got together and manipulated the system by keeping Tim Tebow off their ballots entirely (what he's not even the third best college player?) to ensure that one of the Big 12 guys won ...
In the method used for the Heisman voting (it's a variation on the Borda count), this strategy is useful. It's called "burying" your strongest opponent. The Heisman committee has been using this method for a while and voters are used to it, so I find this story quite plausible.
The recorded votes reflect a mixture of true preferences and strategic thinking -- just as the results of plurality elections reflect the strategic choices of some voters to support their lesser evil rather than their true preference. For this reason, I don't think you can really predict that Bradford would have won an instant runoff election. The pressure to vote strategically would have been both different and far, far less. I suspect that would have resulted in a different mix of recorded rankings.
It's difficult to infer the results under method B from the results under method A, except in special cases.
Dec 15, '08
Re: Paddy - the person with the most votes should win.
In the Heisman voting there were a total of 925 votes. If you give the victory to the person with the most first place votes then Tebow would have won with 33.4% of the total votes vs Bradford's 32.4% and McCoy's 28.7%. I don't find that very satisfactory.
And if you total the votes each of them got for either first, second or third Bradford had 811, McCoy 784 & Tebow 750.
Maybe the Big 12 voters skewed the vote but that's a different issue.
Dec 15, '08
Dang, posted too quickly. Re: 811, 784, 750 - That means Tebow wasn't named at all on 175 ballots, McCoy on 141 and Bradford on 114. Again, Bradford was more popular.
Dec 15, '08
Bob Richard,
You say that with IRV compared to Borda, "The pressure to vote strategically would have been both different and far, far less."
How do you figure that "far less" claim? Do you have any expected utility calculations? (I.e. probability that a given strategy will work times utility payoff).
I suspect you actually have no idea what you're talking about.
Dec 16, '08
An election? Last weekend? Yup. The Heisman Trophy - the most prestigious individual honor in all of sport - is awarded through a vote of 870 sports journalists and 55 former winners.
So, you're saying you think OJ is a good elector?
Seriously, it's a good comparison, imo. The character of the ensuing discussion indicates, perhaps, why coming up with a better model might be the easiest bit.
Dec 16, '08
This is just more smoke and mirrors from the pushers of this unconstitutional vote ranking scheme. They must have alot of money - or friends in the media, to be able to keep spewing out this bilge.
The Heisman Trophy isn't a vote, it's a measurement of qualification for the award. To try to make an anology between the Heisman and an election is sheer idiocy. The award goes to the best player, not the most preferred or popular.
IRV people constantly claim that it would assure that election winners would be "preferred by a majority of voters." Yet in practically every case, unless a candidate wins a majority of FIRST CHOICES, the end result is still a second choice candidate.
The pro IRV arguments are false and disingenuous. They claim it will help 3rd Parties, yet they also claim it will eliminate the 3rd Party spoiler effect. Huh? How can it help "3rd Parties" if it is designed to eliminate their influence?
IRV wouldn't even be a good choice for picking America's next top model, because even something as frivolous as that has certain qualifications that must be met. IRV solves nothing. It just adds another layer of confusion to a process that is usually pretty straightforward: whoever gets the most votes wins.
Dec 16, '08
It still doesn't negate the fact that Texas doesn't fall into the ocean because O.U. sucks.
9:42 a.m.
Dec 16, '08
whoever gets the most votes wins.
Matt, how do you determine that?
In my ice-cream example above in the comments, are you really arguing that when 70% of the people have strawberry as their least-favored choice, then the preferred outcome for the group is strawberry?
Dec 16, '08
One of the forces that leads to the inclusion of 3rd party concerns by a major party is the spoiler effect, it is payment of blackmail, essentially, but is that outcome negative? The Party closest in view to the 3rd finds it to its advantage to pay attention.
Unless you propose that said 3rd party would garner a pile of #2 votes if it were ignored, they lose that leverage since its #1 vote would be low and present no threat. If there is no threat to the politically nearest chances of winning there is little political leverage.
There have been issues that either Party would ignore if not threatened in an election. Gordon Smith's fake moderate position cost him the 3rd party votes while gaining him some Democratic voters, it was a choice he made in his public presentation. I doubt the next (R) will ignore that piece, but that also means the next (R) will have less incentive to play at being moderate when not so.
It might be interesting to run a large enough state wide poll with IRV to see how that would have shaken out. I've run it through my head regarding 2CD and I'm not coming up with anything I'd be comfortable presenting as an analysis.
Dec 16, '08
Hey I know of a really cool website called the voting site . The sites mission is to evangelize the instant run off voting. Users can create there own instant run off elections and vote in other peoples instant run off elections. The site does a good job of showing how instant run off voting works. It also has facebook integration.
Dec 16, '08
Posted by: Garrett | Dec 16, 2008 9:23:40 AM
It still doesn't negate the fact that Texas doesn't fall into the ocean because O.U. sucks.
Did you hear about the Aggie that moved to OK? Raised the average IQ of both states!
Or the Aggie that went to Harvard? Couldn't find the Chemistry lab on the first day, and asked a passer-by, "Do you know where the Chemistry lab is at?" The Harvard guy replied, "At Harvard, we don't finish a sentence with a preposition." To which the Aggie replied, "OK. Do you know where the Chemistry lab is at, asshole?"
Dec 16, '08
<the heisman="" trophy="" -="" the="" most="" prestigious="" individual="" honor="" in="" all="" of="" sport="">
The World Cup's Golden Boot is "numero uno" everywhere else in the world KC.
Dec 16, '08
Of course, the system itself affects the vote. How many Big 12 voters who buried Tebow wouldn't have done so if they weren't afraid of their second- or third-place votes giving him the overall lead? If the Heisman voters could vote only for one, would the anti-Tebow forces have been able to coalesce behind Bradford or McCoy sufficiently? If they could select four players, would Graham Harrell have split the vote further? Gaming the system will always happen -- if an IRV or other non-first-past-the-post system started gaining traction in the political world, just look for the Republicans to start propping up the Libertarians to drain off the second-place votes, and the Democrats to do the same with the Greens.
(Personally, I think it's all horseshit anyway, since Heisman voting doesn't take bowl games into account. Vince Young should have won over Reggie Bush in 2005, because he damn well picked up his team and carried them to the championship, while Bush choked in the biggest game of his career. Ditto Troy Smith in 2006 when he got beat like a drum, and let's face it, Tebow's was for winning the title the year before as much as it was for his 9-4 2007 season.)
((Also, you're now quoted by name in the Wikipedia entry on the Heisman.))
Dec 16, '08
Well, as long as your are limiting your comparison to the Heisman trophy.
Even though I'm not a fan of "futbol americano", I know enough to curse the national championship process (echoing the wrath of husband, father, etc.) Perhaps we could compare the electoral college system to the National Champion selection process?
Dec 18, '08
No Kari, I am not suggesting that "when 70% of the people have strawberry as their least-favored choice, then the preferred outcome for the group is strawberry."
That's the phony anti-plurality argument IRV advocates constantly use to obfuscate the issue. Obviously, anytime you have a plurality, the winner (the one with the most votes) will have less than majority support, so the point is moot. IRV will not sove the majority issue because the end result is still merely a fabricated majority - many times its just a matter of recalculating the percentages by treating exausted ballots as if those votes were never cast. This just happened in a North Carolina schoolboard election.
IRV also has the flaw of only counting the ballots of candidates who were eliminated. So some voters get more cracks at it than others. It's the First Choices that denote the will of the public. Second choice votes are just that...second choices.
And for those who say it will help the smaller parties, consider this: The promoters of IRV argue that it will help "3rd Parties" and that it will eliminate the so-called "3rd Party spoiler effect." My question is how can IRV help "3rd Parties" when it is designed to eliminate their effect? It can't. IRV is a cure worse than the disease.
Dec 18, '08
This is not correct. If no candidate obtains a majority in any given round, the candidate with the least support is dropped and ALL the ballots are retallied, with the votes going to the top remaining choices on each ballot (so if your first choice is still alive, then it goes there; if not, it goes to your next choice, or the one after if your next choice has been eliminated).
As for whether IRV helps or hurts minor parties, it's something of a somewhat-informed guess, but a guess nonetheless. It's hard to say with certainty because we have to dominance of the two majors because plurality winner-take-all tends to create a duopoly wherever it's used. There are good arguments that minor parties will not be hurt by eliminating concerns about "spoiling." On the other hand, as you say, under plurality rules, most power that minor parties have is from the threat of spoiling.
The real question is not whether it helps minor parties, but does it help voters. And there the answer is much more clear: Yes. Voters who are happy with one of the big parties no longer need to wage war on their closest ideological cousins, which is the way it works now. And voters who prefer a minor party no longer have to worry that their sincerely cast vote will help elect the greater-of-two-evils, because they can rank their small party choices ahead of their closest big-party candidate (the lesser of evils as they see it).
Dec 23, '08
Posted by: Zarathustra | Dec 16, 2008 6:36:11 AM
An election? Last weekend? Yup. The Heisman Trophy - the most prestigious individual honor in all of sport - is awarded through a vote of 870 sports journalists and 55 former winners.
So, you're saying you think OJ is a good elector?
I can't believe I said that and didn't get an 'SC dig in! Well, from the online dickipedia.com,
As a lifelong dick who excelled at football, Simpson naturally found his way to a school known for both of these pursuits: the University of Southern California. At USC Simpson became a star running back and won the Heisman Trophy in 1968. No Heisman winner has won by a larger margin of victory, nor has one ever gone on to be accused of murdering his ex-wife and Ron Goldman.
That's OK, Kari. It's not the size that counts.
<hr/>