Another Glimpse at Oscar Voting Thanks to the WSJ
Brace yourselves!
Following the announcement of the 2009 Oscar nominees I posted an article talking about Oscar voting and the peculiarities involved with the "preferential tabulation process" or "instant runoff" as it is referred to by Carl Bialik at the Wall Street Journal as he adds his displeasure with the current voting system used by the Academy Awards.
His argument comes complete with graphics, scenarios, eight different ways of voting with several different outcomes and quotes from people such as Michel Balinski, a professor of research at École Polytechnique in Palaiseau, France, calling it "crazy" and Steven Brams, professor of politics at New York University saying he thinks the final vote for the Oscar winner may be even worse than the selection of nominees.
His article is rather in depth so I am going to cherry pick a few items and let you read the rest at your leisure. Let me give you Bialik's graphic followed by a quote for you to ponder.
Brace yourselves for "Ishtar" defeating "The Godfather." Suppose 49 voters award "The Godfather" six points and "Ishtar" only four. One voter grants the desert debacle four points and the mafia masterpiece three, and the remaining 49 award "The Godfather" three points and "Ishtar" only one point. "Ishtar" actually wins with a median score of four points compared to "The Godfather's" three points. Prof. Balinski, in turn, calls range voting a "ridiculous method," because it can be manipulated by strategic voters.
Of course this is looking at an extreme situation and the article goes on to pose reasons as to why perhaps Rocky beat the likes of Network and Taxi Driver for Best Picture, but as interesting as it is to learn the intricate details of the voting system the fact of the matter is that no voting system will be entirely perfect and bitching and moaning will ensue either way (present company included).
Links from Other Sites You May Like
Showing 10 Comments
~ PLEASE NOTE ~
If, in any way, your comment is an attack on the author of this post or a previous commenter, your comment will be deleted without question.
Add a New Comment |
Click to Read Our Commenting Rules & Guidelines
The voting system does look questionable. Though the Academy has been adopting it since start, it may not necessarily be the most comprehensive, let alone updated.
It only shows how irrelevant the Academy has become, sticking to its rigidity while setting silly rules such as no same actor in one category, embarrassing selections especially in the Foreign Language Film category, and many more.
The Oscars is supposed to be the grandfather of awards. Take a hint from BAFTA or even the Globes.
How about if everyone just voted for their favourite film of the year? No seconds, thirds etc. – just their favourite. Determining the five nominees would simply be a case of taking the five highest totals. Once the nominees have been determined there would then be a second round of voting – voters would then vote for their favourite film from the five nominees. Again the winner would simply be the one with the most votes. The only fault I can see with this system is there may not be five different films in total voted for – I don't think this is very likely though. Much more likely is there being a large 'drop-off' from the most popular film to the second most, to the third most and so on. However I don't see that as anything other than a true reflection of how popular a given film is amongst voters.
Note that the most unfair system is what's done to win the Oscars — a "first past the post" approach where a nominee theoritically could win even if 79% of voters might see it as the worst choice. But this allows for those fun "surprises" on election night.
The single transferable vote system (not really "instaant runoff" which is winner-take-all rather than proportional representation) used for nominees is actually quite a fair system for ensuring a good reflection of Academy voters in each category. Winning a nomination is a win in itself, and this ensures that a full range of Academy voters get to help elect a "winning" nominee.
Note that the best song system is like the range voting system Balinski derides. Voters in that categroy rate each song. It means you don't get nearly as much diversity of opinion (witness two of three nominated songs ffrom same movie, say). It also means people who don't like Bruce Springsteen could have given his Wrestler song very low ratings in a strategic way.
The Academy should stick with "STV" and change the song nomination system.
Just one question. I know that voters vote on their respective categories on the nomination ballots. Once the nominees are selected, do they still vote only for their category and best picture on the final ballot? Or do they get to vote for all?
@cmac: That would be a total nightmare because it would mean just about film receiving about 600 votes would practically guaranteed a nomination even if the Academy's other 5,000+ members hated it. All it would take is a studio strong arming the right people and boom, instant nomination. It would also mean all 6,000 members could think a movie was the first or second best film of the year, but have it not get enough votes to be nominated since you're only voting for one film. I still think the Academy needs to go back to its old system of ten nominees with voters voting for the top five. It makes finishing anywhere on a ballot important and virtually eliminates snubs.
The Academy has used the current single transferable vote nomination system for all the big categories since 1936, so I'd be interested in hearing more about what CMAC means. Was there a different "voting for the winner" system with 10 nominees for awhile?
@malevolentmuse: I disagree – Only five films can end up being nominated. If the first film receives say 1000 votes, the next receives 800, the next 700, the next 650, the next 600 then those are the films that get nominated. The remaining 1250 votes don't count, even if 599 of them are for the same film. Similarly, if the top five amounts are say 650, 650, 640, 610, 610 then those are the films nominated, and the remaining 1940 votes don't count, even if 609 of them are for the same film. I don't really see why this wouldn't be an effective and fair system. Maybe I'm missing something.
@cmac: You're focusing on the wrong aspect of what I said. If voters only vote for one film, but everyone agrees one the second or third film those films can be left off giving every film with 600 votes a shot when a film everybody agrees is one of the five best gets left out. The argument against the current voting system is it puts too much emphasis on films getting first place votes. You compound that problem ten-fold by eliminating everything except films to get first place votes. You don't solve the problem by doing even more of what's causing the problem in the first place.
@malevolentmuse: Okay – I definitely see your point, and I agree with you in principal, but I'm not sure I agree that there shouldn't be emphasis on first choice films at all – I think there's room for this to work. If you'll indulge me I'm going to run something by you to try and see if my theory holds water. This would be a two round process. The first round voters vote for their favourite film of the year. The top scoring films become the years nominees. Then the voters choose which of the nominees is their favourite and vote again to decide the outright winner. If we assume a voting pool of 6000 members, let's look at a hypothetical version of this years lineup:
1200 vote for Slumdog Millionaire
1000 vote for Benjamin Button
600 vote for Frost/Nixon
600 vote for Milk
599 vote for The Reader
598 vote for Revolutionary Road
592 vote for The Wrestler
260 vote for The Dark Knight
200 vote for Wall-E
150 vote for Doubt
100 vote for In Bruges
90 vote for Frozen River
9 vote for Australia
2 vote for The Love Guru
In this case the nominees would be identical to this years actual nominees, because every vote for any film from Revolutionary Road right down to The Love Guru is discarded, leaving the 5 most voted for films. In the above example Revolutionary Road misses out by a single vote but it's still fair because it's a straight vote count – no fancy manipulations. Now the interesting scenario would be if one of those Love Guru fans decided that, as great as The Love Guru was it wasn't actually as great as Revolutionary Road after all and instead cast their vote for that instead. This would result in Revolutionary Road and The Reader both having 599 votes and The Love Guru only having 1 (which of course is irrelevant). How then would it be decided which of the joint 5th films be entered as an official nominee? I think it might be fun to put both of them in and in this case there would be 6 best picture nominees. And if The Wrestler picked up another 7 votes from the bottom of the heap then it would be entered aswell and there would be 7 nominees. Remember though that this would only happen if films around the cut-off point have exactly the same number of votes. I'm no statistical expert or anything but it seems to me that this isn't that likely to happen with a pool of only about 6000 voters. The more I think about this idea the more it sounds like fun! I suppose when the nominees are decided (however many that ends up being) you still have the same problem in the final round of voting ie. there still has to be some way of breaking a potential tie when you get right down to it. Maybe in the second (final) round of voting voters must rate the nominees from first choice to last – the 2nd, 3rd, 4th etc choices only being counted in the event of a tie. For example say in the final round both Slumdog Millionaire and Benjamin Button receive exactly 2000 votes each with the other nominees receiving the remaining 2000 between them in varying amounts. Then, and only then will the number of 2nd choice votes for Slumdog and Button be counted. If those totals are also identical then you would look at how many 3rd choice votes each received and so on. This sounds like it could work, and also be fair. Of course it's after 2am here right now so I may look at this tomorrow in horror, but it was fun getting it down on paper, as it were.
Folks, The current system does make getting first choice rankings important, but most nominees also likely had to be the second and lower choices of supporters of other potential nominees. Single transferable vote does a great job at balancing those values, with the result being a slate of nominees that really relect different views and opinions well.
Votes aren't "wasted" by teh way — if your first choice can't win, your ballot goes to your second choice. See a video on how the system works here:
http://www.citizensassembly.bc.ca/flash/bc-stv-count