Electing a 21st Century President
Steven Kates is an Australian economist (a very good one - his recent book is well worth reading, and if you can't wait for Amazon and don't like the price tag, here's something of a summary by him published last month in Quadrant, an Australian magazine).
So he's a serious thinker. But I don't know about this piece he's recently posted online at Quadrant. His point is captured in this paragraph:
Computer chess programs can now beat the human world champions almost ten times out of ten. They never forget, they never make mistakes and they never get tired. They can also see a hundred moves ahead. The Obama campaign is now run by people with an astonishing ability to analyse an immense quantum of social data through the most upscale and modern statistical techniques. Through the use of those statistical programs they are able to design a campaign that will zero in on the most effective campaign structure and message that will draw the highest number of votes. In this way they too can see, politically speaking, a hundred moves ahead.
Everyone here is a lot closer to this than me (and, possibly, Steven). Is he right?
- Comment (29)
- · Quote
- · UnfollowFollow (4)
- Pages:
- 1
- 2
- Pages:
- 1
- 2
Comments :
May '11
Re: Electing a 21st Century President
Two thoughts:
1) Mr. Kates seems to be imbuing Mr. Obama and his team with supernatural powers. Is this not akin to superstitious, magical thinking that turns an enemy into Satan incarnate? Mr. Obama is "just" a man who makes mistakes like everyone else. (He's also a lousy President but that's another issue. Or maybe not.)
2) If Mr. Kates is right, I sure hope whoever runs against him has a better computer and better operators.
Nov '10
Re: Electing a 21st Century President
I don't think so. As complicated as chess can be, it has hard rules. People are pure chaos. Trying to predict what one person does is nearly impossible. Trying to predict how a nation will react is chaos to the 300 millionth power. It can't be done.
Feb '11
Re: Electing a 21st Century President
No he is not right.
I remember an anecdote I read about the Vietnam war. The story was that at some point long after the war began Robert McNamera- really his staff, no doubt- fed into a computer all the data about the relative strengths of the US and North Vietnam, and McNamera eagerly awaited the answer. After the calculation was complete the answer came: The US won in 1964.
Oops. So color me skeptical that the sort of calculation as described by Steven Kates actually means much, especially when he is faced with a candidate such as Perry who also seems to capable of playing the same game.
Oct '10
Re: Electing a 21st Century President
What utter and silly nonsense. Does this guy have any concept of how big a magic act he is waiting for the curtain to rise for???
Obama and his administration still could be replaced by the first 400 names in the Boston phone directory and America could be saved.
Jan '11
Re: Electing a 21st Century President
My instinctive distrust of the "solve all the world's problems with an equation" approach (refer the climate models) says to me that you're probably right - and a feeling that Obama's team isn't as smart as they think they are (refer Paul Krugman). But I think there is something in looking deeper into the data analysis and trying to use it to test some of the outcomes against the anomolies that don't fit the existing "rules of thumb" as he calls them.
Jul '11
Re: Electing a 21st Century President
Raycon summed it up nicely - "what utter and silly nonsense". Let's, for the sake of argument, say voters' reactions were predictable enough to make such a precisely orchestrated campaign possible. Campaigns are at the mercy of unforeseen events (e.g., financial meltdowns in September), gaffes by the candidates, scandals, etc. What sways many voters' minds is not what the campaign set out do to (the sales jobs), but how the candidates respond to the unexpected. Obama looked calm and presidential in Fall 2008 when McCain wigged out and "suspended" his campaign in response to the financial crisis (although we now know, that his calmness is his perpetual fear of decision-making).
Finally, let's get real about who these guys are. Software capable to predict 100 moves out is still subject to the "Garbage In - Garbage Out" Rule - Obama's team is so far removed from middle America they are incapable of understanding just how unattractive their own ideas are ... Obamacare anyone? More stimulus? Israel-bashing? They think these are good things and we can only hope they plan their campaign with that in mind.
Sep '10
Re: Electing a 21st Century President
raycon:
Obama and his administration still could be replaced by the first 400 names in the Boston phone directory and America could be saved. · Aug 31 at 9:01pm
True, but unfortunately if he is replaced he will be replaced with a Republican which means we are in trouble.
Re: Electing a 21st Century President
This is why I'm increasingly hesitant to pronounce upon America's future. When I was home last, I noticed this and realized how deeply skewed my impressions are. I don't live there and I read what interests me--and it is, apparently, not quite what interests most people in America. As for predicting voter behavior, well, these two guys at Yale are not the only ones who have tried.
Edited on Sep 1, 2011 at 5:17amDec '10
Re: Electing a 21st Century President
If they could control most or all of the variables they might begin to develop a degree of predictive or manipulative accuracy. They cannot control the variables and therefore, are not even close to the degree of political power ascribed by the quoted article. The power base of the Democrats is the constituency created by taking money from some and distributing it to others and the tag-a-longs who want to be part of the redistributive process. It is that simple!
Jun '10
Re: Electing a 21st Century President
So... Republicans don't have stats people too... has this guy ever even heard of Mike Murphy?!?!
The more relevant predictive model is the model the current administration is running for this country. I might just be a poor country conservative that can't use a computer, but it seems to me that THAT model has shown... disappointing results....
May '10
Re: Electing a 21st Century President
Team Obama's data-sorting was very useful in the 2008 primary, in spotting the delegate opportunities available during the lesser caucuses, where Activists Walk the Earth. They were helped along in this endeavor by the fact that Hillary forgot to hire smart people. Oops.
Organizing for America has indeed compiled a whole lot of data, but so far has not been effective at exploiting it. Their pinpoint strategies so far have made Obama sound like Cleverbot. Maybe they're waiting for the Big Reveal next year, when Obama unveils his Fully Operational Battle Station, eliciting gasps of panic from the cocksure rebel forces.
But given his track record, I doubt it. His big surprises are neither big nor surprising. Perhaps these tools could be used as a force multiplier for a salesman with an actual product to sell. Obama doesn't have one.
PS by "these tools" I meant Organizing for America.
Edited on Sep 1, 2011 at 5:53amOct '10
Re: Electing a 21st Century President
It is true that "Barack Obama is the short priced favourite to win the Presidential election in 2012."
It is true that "[Obama] also has the overwhelming support of the mainstream media which will suppress [his] negatives and highlight every positive."
It is not true that "Computer chess programs see a hundred moves ahead." Computer chess is hard, and relies on clever science, implementation tricks to squeeze out performance, and a bunch of (human generated) rules-of-thumb derived from experience.
It is true that mining terabytes of data in clever ways can give insights into consumer behaviour. But only on the margin. WalMart has a famously sophisticated data-mining operation with vast amounts of data and the ability to run experiments with real consumers spending their own money day after day. And they still get things wrong.
Applying these techniques to a situation where the consumer only undertakes one meaningful transaction (i.e. only votes for real once - we all know the difficulties of voting intention polls) seems ... challenging.
If there are "predictive Modeling/Data Mining’ specialists" who can analyze the Twitterverse "to inform campaign strategy and critical decisions" in a meaningful manner, I would be very surprised.
Jul '10
Re: Electing a 21st Century President
Is this the same computer program that was predicting "global warming" with a hockey stick?
Dec '10
Re: Electing a 21st Century President
What about Obama's performance thus far has Steven Kates thinking Obama is competent? The one single thing that has astonished people the most about the Obama administration is not the arrogance or the radicaliness of it. Its the sheer utter incompetence. From the Russian "reset" button flub to this latest rebuttal by Boehner forcing the President to reschedule his speech, amateur hour has been the name of the game since he took the oath of office.
I would even go so far to argue that even his original campaign wouldn't have been anything special if the press core had completely and totally fallen in love with the man.
PS
Chess an other similar basic pattern recognition games where AI out performs humans. Take it from a gamer, in any situation where actual strategy is involved Humans stomp on AI so hard that as a Computer Scientist its almost depressing...almost.
Jan '11
Re: Electing a 21st Century President
Data mining on grocery shopping is one thing, but political campaigns are different.
Obama's in a much different position now. In 2008, he could discover whatever appealed to the voter (centrism, rising above politics, etc.) and then craft his statements to exploit those poll-tested phrases. Back then, however, he had no record to compare against. Now he does. A phrase that "worked" in 2008 won't work now, no matter what the poll says about its appeal to the voter.
Feb '11
Re: Electing a 21st Century President
There is a long history of rather breathless predictions about the impact of computer technology, in politics as in other things. There's a pretty good 1964 novel by Eugene Burdick, "The 480," about the exploitation of a computer model (running on the mighty IBM 7094!) to promote a dark-horse candidate. The title of the book is taken from a categorization of the American electorate developed (in real life) by Simulmatics Corporation in support of John F Kennedy's campaign.
Jul '11
Re: Electing a 21st Century President
One of my favorite maxims is "When you don't know what to think, look." An empirical look at these bozos reveals that whatever data they're mining, it's not working. If they were so smart, we'd not be in the mess we're in.
I don't think the problem is how smart they are, but rather how dumb we are. Claire noted how insular our thinking is; we speak to each other and the world looks rational, but we're not our whole world. Too much of the rest is a place where American citizenship isn't worth paying for, where gazillions of folks can't find Mexico on a map.
A simple fact is that despite the desperation into which we are descending, being pushed by the Obamanoids to even greater rates of descent, millions of our fellow citizens think this is just peachy, and approve of it all. That just blows my mind. Millions.
We're in a mess for sure, but it's not because those guys mine data better than we do. It's because our bill has come due. It'll be a dangerous and painful payment.
Feb '11
Re: Electing a 21st Century President
A year or so ago there was an interesting article, which I now can't find, about the use of software to optimize electoral redistricting in favor of particular parties. IIRC, the Dems were making more & better use of such SW than the Repubs.
Apr '11
Re: Electing a 21st Century President
If Obama's team is predicting 100 moves ahead of their opponents, how do we explain his recent spate of poor decisions....10 days in Martha's Vineyard, the silly State of the Union Address incident, his latest strategy to demonize congress? His campaign seems like it is floundering to me.
Regarding the idea the average citizen is not paying attention to ins and outs of politics, I think they are aware that a lot of people are out of work, prices are rising and we are involved in a third war. They may not be following who is calling who a terrorist and how big the deficit acutally is, but they know things are not good.
We can't afford to be complacent, but this next election has the potential to be landslide.
Dec '10
Re: Electing a 21st Century President
Numerical and statistical models can't make up for the human touch - the one element Obama will never understand. Without this understanding statistical models are useless and prone to significant errors of interpretation. I don't need a statistical model to make this prediction which will be correct 99 times out of 100: Obama won't get 40% of the vote in 2012.