Dilbert creator Scott Adams, who has a fantastic rationalist-compatible blog, is giving Donald Trump a 98% of becoming president because Trump is using advanced persuasion techniques. We probably shouldn’t get into whether Trump should be president, but do you think Adams is correct, especially about what he writes here. See also this, this, and this.
It was because of Nate Silver’s track record that I initially had high confidence in his estimate. Then as I read his justification my confidence in his estimate decreased. I think he’s just being lazy in his justification, here, when he says things like:
So, how do I wind up with that 2 percent estimate of Trump’s nomination chances? It’s what you get if you assume he has a 50 percent chance of surviving each subsequent stage of the gantlet.
To be fair to Silver, when he wrote the article he might not have considered Trump’s campaign plausible enough to give serious thought. I suspect that if Trump continues to perform well in the polls Silver will give a more thoughtful and realistic analysis later on.
Were any of Silver’s previous predictions generated by making a list of possibilities, assuming each was a coin flip, multiplying 2^N, and rounding? I get the impression that he’s not exactly employing his full statistical toolkit here.
What Adams does is that he looks at Silver’s estimate, says that it is way too low and then takes 1 minus Silver’s estimate as his own estimate just to make a point. He does not attempt any statistical analysis and the 98% figure should not be taken seriously.
What Adams has said he’s doing is simulating the future along the mainline prediction—i.e. nothing too weird happens—and under his model, Trump is guaranteed to win. Then he says “well, maybe something weird will happen” and drops that confidence by 2%, instead of a more reasonable 30% (or 50%).
Does Adams have a track record at predicting this sort of thing? I am not aware of any instances he’s said “here is a master persuader trying to do X, they will succeed” and them having failed, but I can’t remember more than one instance of him saying that and it being correct (and I don’t remember the specifics), but I don’t follow Adams closely enough to have a good count.
I think that Adams is raising the sort of challenge that Silver is weakest against: Trump’s tactics are a “black swan” in the technical sense that no candidate in Silver’s dataset has run with a similar methodology. That Silver thinks Herman Cain’s campaign is the right reference class for Trump’s campaign seems to me like a very strong argument for Silver not getting what’s going on.
Why do so many people see Adams as being rationality-compatible? I’ve seen very little that he has to say that sounds at all rational or helpful. Cynical != rational.
Having written a rationality-compatible book isn’t the same thing as writing a rationality-compatible blog. (It surely indicates being able to write a rationality-compatible blog, but his actual goals may be different.)
Well… Scott Adams has a lot of money. I am willing to bet that Trump will NOT become president, at EVEN ODDS. Scott, if you read this, how about a wager? I propose a $10,000 stake.
Despite his frequent comments that he’s “betting” on Trump and that Silver is “betting” against Trump, Adams’s position is that gambling is illegal when pressed to actually bet. This means one of the big feedback mechanisms preventing outlandish probabilities is not there, so don’t take his stated probabilities as the stated numbers.
(In general, remember how terrible people are at calibration: a 98% chance probably corresponds to about a 70% chance in actuality, if Adams is an expert in the relevant field.)
Despite his frequent comments that he’s “betting” on Trump and that Silver is “betting” against Trump, Adams’s position is that gambling is illegal when pressed to actually bet.
And Adams himself says the “smart money” is on Silver’s prediction! I think Adams’s prediction is more performative than prognostic, even allowing for ordinary unconsciously bad calibration.
Forgetting what I know (or think I know) about Scott Adams, Donald Trump, Nate Silver, Jeb Bush, whoever, and going straight to the generic reference class forecast — I’m very sceptical someone could predict US presidential elections with 98% accuracy 14 months in advance.
Well, he’s very likely substantially healthier than the average 69-year-old American man, so I’d be willing to bet at 1⁄50 odds that he will survive to the election.
I think Scott Adams wildly overestimate the power of conversational hypnosis. First of all, yes, there have been prominent public figures who are well versed in the art. But that’s no argument at all: how many people are trained in conversational hypnosis (or NLP, or what have you), and how many of those are hyper-successful? And how many hyper-successful people are not trained in Ericksonian hypnosis? You could even make the point that Steve Jobs and Bill Clinton were successful despite being trained in that art.
There’s also something to be said about linear return on persuasion. If you are 2X more persuasive than your opponent, would you gain twice the supporter? I’m not very confident in this hypothesis too.
There might be a network externality effect with persuasion, where the more people I persuade the more persuasive I become because of social proof issues. In this situation, the returns to persuasion are exponential.
I think Adams is right that Trump has played the media exceedingly well and he has clearly surprised a lot of people. Some Republican pollsters have focus-grouped Trump supporters and found an extreme level of antipathy among them toward “establishment” Republicans. So it is unlikely his current supporters will abandon him in a sudden collapse, which is the failure mode a lot of Trump-skeptics have been describing. That means Trump will likely stay in the race for a long time—unless he gets bored and drops out. I doubt Trump will actually drop out though, he seems to enjoy the fray and clearly hates many establishment conservatives enough to stay in just to have a platform to keep attacking them.
Most likely Trump will split the anti-establishment vote with Ben Carson and eventually most of the establishment candidates will drop out and throw their support to an establishment survivor, who will manage to beat Trump with solid but not huge majorities and take the nomination. If Trump does manage to win the nomination, it is unlikely he will win the white house—odds are less than even, maybe 2:1 against him. Overall I would estimate a ~10% chance Trump wins the presidency.
Dilbert creator Scott Adams, who has a fantastic rationalist-compatible blog, is giving Donald Trump a 98% of becoming president because Trump is using advanced persuasion techniques. We probably shouldn’t get into whether Trump should be president, but do you think Adams is correct, especially about what he writes here. See also this, this, and this.
I think Scott Adams has taken to trolling the readers of his blog.
Taken to? He’s been doing it for like a decade at this point.
I wouldn’t put it at 98%, but I definitely wouldn’t put it at Nate Silver’s 2%, which I think comes from an analysis that is just way too simplistic.
I would take Silver’s analysis over Adams’ any day. Look at their respective prediction track records.
It was because of Nate Silver’s track record that I initially had high confidence in his estimate. Then as I read his justification my confidence in his estimate decreased. I think he’s just being lazy in his justification, here, when he says things like:
To be fair to Silver, when he wrote the article he might not have considered Trump’s campaign plausible enough to give serious thought. I suspect that if Trump continues to perform well in the polls Silver will give a more thoughtful and realistic analysis later on.
Were any of Silver’s previous predictions generated by making a list of possibilities, assuming each was a coin flip, multiplying 2^N, and rounding? I get the impression that he’s not exactly employing his full statistical toolkit here.
Isolated demands for rigor—what do you think Adams is doing? (I think he’s generating traffic.)
But sure, I agree, that’s more of a reasonable prior than an argument. There’s more info on the table now.
What Adams does is that he looks at Silver’s estimate, says that it is way too low and then takes 1 minus Silver’s estimate as his own estimate just to make a point. He does not attempt any statistical analysis and the 98% figure should not be taken seriously.
What Adams has said he’s doing is simulating the future along the mainline prediction—i.e. nothing too weird happens—and under his model, Trump is guaranteed to win. Then he says “well, maybe something weird will happen” and drops that confidence by 2%, instead of a more reasonable 30% (or 50%).
Does Adams have a track record at predicting this sort of thing? I am not aware of any instances he’s said “here is a master persuader trying to do X, they will succeed” and them having failed, but I can’t remember more than one instance of him saying that and it being correct (and I don’t remember the specifics), but I don’t follow Adams closely enough to have a good count.
I think that Adams is raising the sort of challenge that Silver is weakest against: Trump’s tactics are a “black swan” in the technical sense that no candidate in Silver’s dataset has run with a similar methodology. That Silver thinks Herman Cain’s campaign is the right reference class for Trump’s campaign seems to me like a very strong argument for Silver not getting what’s going on.
He has an excellent track record of saying outrageous things—that’s what he is optimizing for, I think.
Why do so many people see Adams as being rationality-compatible? I’ve seen very little that he has to say that sounds at all rational or helpful. Cynical != rational.
See my review of his book: http://lesswrong.com/lw/jdr/review_of_scott_adams_how_to_fail_at_almost/
Having written a rationality-compatible book isn’t the same thing as writing a rationality-compatible blog. (It surely indicates being able to write a rationality-compatible blog, but his actual goals may be different.)
Well… Scott Adams has a lot of money. I am willing to bet that Trump will NOT become president, at EVEN ODDS. Scott, if you read this, how about a wager? I propose a $10,000 stake.
Despite his frequent comments that he’s “betting” on Trump and that Silver is “betting” against Trump, Adams’s position is that gambling is illegal when pressed to actually bet. This means one of the big feedback mechanisms preventing outlandish probabilities is not there, so don’t take his stated probabilities as the stated numbers.
(In general, remember how terrible people are at calibration: a 98% chance probably corresponds to about a 70% chance in actuality, if Adams is an expert in the relevant field.)
How convenient for him.
And Adams himself says the “smart money” is on Silver’s prediction! I think Adams’s prediction is more performative than prognostic, even allowing for ordinary unconsciously bad calibration.
Forgetting what I know (or think I know) about Scott Adams, Donald Trump, Nate Silver, Jeb Bush, whoever, and going straight to the generic reference class forecast — I’m very sceptical someone could predict US presidential elections with 98% accuracy 14 months in advance.
Actuarial tables give him a roughly 2% chance of dying before the election.
Well, he’s very likely substantially healthier than the average 69-year-old American man, so I’d be willing to bet at 1⁄50 odds that he will survive to the election.
Did Adams praise Obama for skillful use of vagueness? “Hope” seems to be in the same category as “take your country back”.
I think Scott Adams wildly overestimate the power of conversational hypnosis.
First of all, yes, there have been prominent public figures who are well versed in the art. But that’s no argument at all: how many people are trained in conversational hypnosis (or NLP, or what have you), and how many of those are hyper-successful? And how many hyper-successful people are not trained in Ericksonian hypnosis? You could even make the point that Steve Jobs and Bill Clinton were successful despite being trained in that art.
There’s also something to be said about linear return on persuasion. If you are 2X more persuasive than your opponent, would you gain twice the supporter? I’m not very confident in this hypothesis too.
There might be a network externality effect with persuasion, where the more people I persuade the more persuasive I become because of social proof issues. In this situation, the returns to persuasion are exponential.
I think Adams is right that Trump has played the media exceedingly well and he has clearly surprised a lot of people. Some Republican pollsters have focus-grouped Trump supporters and found an extreme level of antipathy among them toward “establishment” Republicans. So it is unlikely his current supporters will abandon him in a sudden collapse, which is the failure mode a lot of Trump-skeptics have been describing. That means Trump will likely stay in the race for a long time—unless he gets bored and drops out. I doubt Trump will actually drop out though, he seems to enjoy the fray and clearly hates many establishment conservatives enough to stay in just to have a platform to keep attacking them.
Most likely Trump will split the anti-establishment vote with Ben Carson and eventually most of the establishment candidates will drop out and throw their support to an establishment survivor, who will manage to beat Trump with solid but not huge majorities and take the nomination. If Trump does manage to win the nomination, it is unlikely he will win the white house—odds are less than even, maybe 2:1 against him. Overall I would estimate a ~10% chance Trump wins the presidency.