Skip to content

Old story: The insiders go after the stat guys. New twist: in politics

Nov 1, 2012, 3:00 PM EDT

Nate Silver

Nate Silver is known for two things: (1) being one of the old school Baseball Prospectus people, where he developed the PECOTA projection system; and (2) being a political polling analyst/prognosticator and proprietor of the blog, which appears in the New York Times. He used to crunch baseball numbers. Now he crunches political polling data. Viva varied interests and transferable skills.

One of the facts of life for the early Baseball Prospectus folks was the skepticism to the point of disdain they received — and in some cases still receive — for their methods, conclusions and tone, with said skepticism and disdain coming most significantly from the establishment baseball press.  The worst of that is long over — most baseball writers now accept that general take — and it’s actually notable now when someone whips out a decade-old criticism of sabermetrics and other Baseball Prospectusy things.

Silver is reliving the bad old days in the political arena, however, as in recent weeks a healthy portion of the political punditry has become consumed with attacking Silver, his methods, his conclusions and his tone. While I don’t have nearly the investment in the political stuff as I do the baseball stuff, as far as I can tell, most of the criticism of Silver’s work is based on (a) a basic misunderstanding of statistics and what they can and cannot prove; and (b) a resentment of sorts that someone from the outside, as opposed to political writers who have made their bones pressing the flesh, is making headlines and earning a paycheck in their business.

Gee, sound familiar?

I am not going to jump into the specifics of that debate here because this is a baseball blog and not a political one, but the broad strokes of it all are instructive for our purposes.  The best summary of it I’ve seen so far — one which actually explains why these camps fight the way they do as opposed to merely arguing up one side or down the other — comes from Mark Coddington, who has a great post up today talking about it all.

The executive summary: when someone gets their information and their authority from being on the inside, they are inevitably wary, and often hostile to those who seek to play in their sandbox without getting their information and authority from the inside themselves.

I would also add that those on the outside have historically tended to be overly dismissive of information from people who work on the inside, and have their own history of hostility toward their inside counterparts. There are differences in how that is all manifested — one side clearly is the establishment here and one side the newcomers, which shapes the rhetoric of it all — but it’s mostly a fight about how one comes by information and what one considers to be legitimate information.

It’s a fascinating topic. One which I think serves all of us who care about the information we get and are critical of its sources. And Coddington, I think, does a great job of laying it all out without getting sucked into the minutiae of the actual warring camps.

  1. schlom - Nov 1, 2012 at 3:08 PM

    Is that really why Silver is being criticized? I thought it was because Republicans thought there were flaws in his models which have overstated President Obama’s re-election chances?

    • philsieg - Nov 1, 2012 at 3:13 PM

      Here’s one of those thoughtful Republicans attacking the models that got 49 of 50 states right in ’08 and nailed the 2010 midterms that were highly favorable to the right. Memo to Dean Chambers: When you got nothin’, attacking a guy on his appearance is rarely a good idea if you’ve lost track of how many chins you have (see photo).

      • aceshigh11 - Nov 1, 2012 at 3:42 PM

        I hope that bloated, bigoted shitstain, Chambers, chokes to death on his own throat fat.

    • Pierre Cruzatte - Nov 1, 2012 at 3:14 PM

      … which is probably total bs, but, yes, I also understood the source of the recent Silver bashing to be a party line thing and not an insider-outsider thing.

      • dowhatifeellike - Nov 1, 2012 at 4:20 PM

        It’s a typical political attack – if you can’t refute the numbers, discredit the source. The problem with that is that Silver has been the best source of political predictions over the last two election cycles.

    • phisticuffs - Nov 1, 2012 at 3:16 PM

      I don’t claim to have a deep understanding of this, but I think the concern is not that the model is flawed, but that the data input into the model (state polls) is itself flawed. Many of the state polls are sampling based on an assumption that Democrats will turn out like they did in 2008 (or in some cases even stronger), when that seems like a dubious assumption (see, e.g., 2010 midterm elections, voter enthusiam gaps, etc.). If the state polls are accurate, Silver’s model should work. If not, then not so much.

      • snowbirdgothic - Nov 1, 2012 at 3:30 PM

        That’s more of a positioned narrative than a concern. The “the other guys are undersampling our guys!” “Oversampling their guys!” narrative is nothing new here, and with many of the polls moving one way it’s time for the other guys to start howling. Silver is the highest-profile bearer of bad news to those guys, so of course he comes under attack.

        And it is worth repeating that the initial charge at Silver came from the guy at “unskewed”, whose methodology is to adjust things until they feel right, and whose assault on Silver consisted largely of a homophobic ad hominem.

      • dowhatifeellike - Nov 1, 2012 at 4:32 PM

        Silver doesn’t just stick to a given system though. He crunches all the numbers in all of the ways they can be crunched, and selects the model that works best. If the national polls were a better indicator than the state polls, he’d use them. Polling has been going on for decades so it’s easy to apply models to past elections and see which worked best.

        Logically, the national polls don’t tell much. Electoral votes cannot be assigned via national polling numbers. Imagine a scenario where one candidate gets 75% of the vote in half the states and 43% in the other half. That candidate will very likely win the popular vote, but in the end the only part that matters is which states he won and which states he lost.

    • Craig Calcaterra - Nov 1, 2012 at 3:21 PM

      I haven’t watched too closely, but yes, there is a fair amount of general media bashing of Silver that isn’t coming from GOPers who think Silver has teh bias. There’s a lot of totally misunderstanding what he’s saying when he says Candidate X has an X% chance of winning.

      • indaburg - Nov 1, 2012 at 3:36 PM

        I think this misunderstanding is reflected in Joe Scarborough’s quote from Mark Coddington: “Nate Silver says this is a 73.6 percent chance that the president is going to win? Nobody in that campaign thinks they have a 73 percent chance — they think they have a 50.1 percent chance of winning. And you talk to the Romney people, it’s the same thing.”

        I think what Nate Silver’s calculations indicated is that the Obama campaign at the time that data was gathered had a 73.6% chance of winning at least 50.1% of the popular vote.

      • stex52 - Nov 1, 2012 at 3:47 PM

        Which goes back to the point that 99+% of the population doesn’t understand statistics. Thank God for HBT.

      • snowbirdgothic - Nov 1, 2012 at 4:24 PM

        What he was saying was that Obama had a 73.6% chance of reaching 271 electoral votes, which is not quite the same thing.

        And much of the “he oversamples state polls” criticism loses its effectiveness if one considers that the race is decided, not in the popular vote, but in the Electoral College, which is of course done state-by-state.

      • indaburg - Nov 1, 2012 at 4:31 PM

        Yes, snowbird, excellent points. The popular vote is not significant when you consider that it is the Electoral College which ultimately determines the winner. It is conceivable that Obama could lose the popular vote yet still win the Electoral College, similar to Bush in 2000.

      • natstowngreg - Nov 1, 2012 at 5:20 PM

        @inda: As a fan of both Silver and Scarborough (having nothing to do with any of the former’s perceived biases, or the latter’s well-known biases), I see where each is coming from. As in the great baseball scout vs. stathead debate, truth lies somewhere between.

        Silver is accountable for his numbers. If he fails this time (not saying he will), the political commentariat will say, “told ya so.” They will proclaim as a genius whomever guessed right this time. Until the next elections, when the commentariat will find some other genius who guessed right.

        In contrast, Scarborough has actual political experience and real-time contacts in both parties. He’s reflecting what’s going on in the campaigns, where they can’t rely on projections. At the same time, he has less accountability, because he isn’t putting specifics out there. After the elections, whether he guessed right or not, he will still have a job, explaining why things happened the way they happened.

        Can’t we all just get along? [Yeah, I know, we can’t.]

        @stex: With all due respect, I see a fair amount of misunderstanding of stats on HBT.

    • Jeremy Fox - Nov 1, 2012 at 4:38 PM

      Well, except that the right hasn’t much gone after, or even much noticed, the other quantitative election forecasting sites like, that give Obama even better odds of winning than Silver does.

  2. Chris Fiorentino - Nov 1, 2012 at 3:19 PM

    Rasmussen Reports is the best online polling website on the planet. I wouldn’t trust a blogger on to be neutral any more than I would trust a blogger on to be neutral. The trick is to find a poll that isn’t biased(or is at least as little biased as is possible) and get your info from there. If you’re a lib, go to the NY Times. If you’re a conservative, go to FoxNews. Then go vote and watch your channel next Tuesday. Unlike a baseball argument…i.e. Trout vs Cabrera…this debate has a winner and a loser. And the winner will be President.

    • skipperxc - Nov 1, 2012 at 3:27 PM

      The trouble with that train of thought is that he tells you exactly how he comes with his numbers and what his sources are: The guy looks at *all* the polls, no matter which way they lean, and then accounts for that lean and how accurate they’ve been in the past. I fail to see the problem with his approach.

      • skipperxc - Nov 1, 2012 at 3:40 PM

        I should clarify that any approach that is as well-documented as his should be subjected to academic criticisms, but, as Craig says, to insist that no candidate is favored to win is the same as saying that the Yankees and the Astros have a 50/50 chance to win one game. It’s true that either could win, but certainly one outcome is more likely than the other. Though admittedly the comparison is less Yankees/Astros and more Yankees/Orioles.

    • philsieg - Nov 1, 2012 at 3:34 PM

      Have you bothered to read anything Silver has done in the past six years? No matter what his personal preferences are, he’s probably the most scrupulously neutral analyst around today. His entire career depends on it. Look at his record. It speaks for itself.

      I know you won’t accept it, but Rasmussen has been shown by Nate in extensive analysis to have a house bias for Republicans (which is no doubt why you prefer it – confirmation bias). It doesn’t meet you quest for an unbiased poll. And he’s also shown PPP to have a house bias for Democrats. As Craig points out, a good deal of this – including your post – comes from a total misunderstanding of what is meant by saying a certain candidate has X% chance of winning either the Electoral College of the popular vote.

      As Nate himself says, nothing he’s written means that Obama has won the election, only that he has a higher probability of doing so that Romney. Dewey defeats Truman – remember?

    • pemcdonough - Nov 1, 2012 at 3:37 PM

      While I agree with the ultimate conclusion of your comment, the idea that Rasmussen is the best online polling website on the planet is kinda laughable. Rasmussen’s polling was eviscerated in the 2010 elections, proving to be off by an average of 5.8%, with a clear electoral bias in favor of the GOP.

      That critique comes from a conservative (albeit libertarian) writer in Doug Mataconis.

    • aceshigh11 - Nov 1, 2012 at 3:39 PM

      Rasmussen is a stooge for the GOP, and you clearly know nothing about Nate Silver’s background or of his rigorous methodology.

      His blog wasn’t even being hosted on back in 2008 when he called 49 of 50 states and all of the Senate races correctly.

      And bravo with the tired, bogus false equivalence of equating the NY Times with Fox News.

      If you actually think the Times is as ideologically biased to the left as Fox is biased to the right (the Times aided Bush into selling the Iraq war to the American people), then you have little understanding of the media and U.S. politics, and probably shouldn’t be voting next Tuesday.

      • Chris Fiorentino - Nov 1, 2012 at 4:17 PM

        LOL Yeah the Times isn’t biased to the left. Next thing you will be telling me that MSNBC, CNN, ABC, CBS, and NBC are also not biased to the left. Whatever.

        You boys want to read your NY Times and call it balanced that’s your call. At least I watch and read Fox with the knowledge that it gives the other side to an otherwise mainstream media that couldn’t be more biased to the left if it was run by people like George Soros…oh wait.

      • snowbirdgothic - Nov 1, 2012 at 4:28 PM

        Hey Chris?

        MSNBC, CNN, ABC, CBS and NBC aren’t biased toward the left.

        (Hey, you asked for it.)

        Also, you may want to look up a little something called “confirmation bias” one of these days. You may find it enlightening.

      • churchoftheperpetuallyoutraged - Nov 1, 2012 at 4:37 PM

        Chris, you honestly think that the NYT is as far to the left as fox news is to the right? When’s the last time the NYT printed blatantly false information, or misled viewers like fox does?

      • philsieg - Nov 1, 2012 at 4:44 PM

        Chris, I’m sorry. I don’t believe you really know what left is. The current edition of the Democratic Party is a center-right construct that has to walk a difficult tightrope between placating their boardroom benefactors and trying to convince their base that they really care about their issues. Look at the actions not the rhetoric. Ronald Reagan would be a Democrat today. He wouldn’t have any choice because the Tea Party Republicans wouldn’t have him.

        As far as the media go, you might get away with calling MSNBC “liberal” in comparison to all other news outlets, although outside the intelligence of Rachel Maddow, the bloviating faux outrage of Chris Matthews and the former right-wing radio host Ed Schultz is all sound and fury, signifying nothing. But to think the rest of your list is liberal just shows how little penetrates the closed information feedback loop that is right-wing news and blogging.

        At last count there are eight independent studies in refereed journals that corroborate that those who get their information from Fox and Fox only are the least informed members of the populace – less informed on a variety of issues than those who watch no cable or broadcast news at all. To put it simple when queried on subjects from the Iraq war to climate change, Fox News watchers get it wrong, consistently and completely (not surprising since a large portion of the right’s strategy is an all-out war on science and knowledge).

        You can google them, but I’m sure if you do you’ll relegate them to the “liberal bias” basement where you seem to cast off everything else you don’t agree with. You need to look up and try to understand confirmation bias, because Fox News depends on it as much as it depends on you not objectively seeking out any balancing or contradictory information to the propaganda that is their mission statement. Until you are willing to take off the blinders, you aren’t qualified to comment on what’s liberal or conservative nor are you informed enough to cast a meaningful vote, no matter for whom you vote.

      • Marty - Nov 1, 2012 at 5:34 PM

        Perhaps the NYT isn’t as biased as Fox or MSNBC, but they haven’t endorsed a Republican for president during my lifetime.

        AND, I believe Nate Silver fostered his political career at the Daily Kos. I think the notion that he could be biased is extremely fucking valid.

      • philsieg - Nov 1, 2012 at 5:44 PM

        Well, in my lifetime the NY Times endorsed Dewey and Eisenhower twice. They stopped endorsing Republicans when the fiscally moderate and socially liberal Republicans were driven gradually from the party in the post-Goldwater purge and during the Reagan ascendency. Since then the Republicans and their surrogates have drawn the political conversation further and further to the extreme right and demanded increasingly more rigid fealty from anyone desiring to wear the label. Where exactly is anyone not part of the extreme right wing supposed to go?

      • Marty - Nov 1, 2012 at 5:55 PM

        So to be clear, you are admitting they are biased, but you agree with the bias so it’s not much of an issue, and because so they should not be considered biased.

      • philsieg - Nov 1, 2012 at 6:13 PM

        I know this is a difficult concept to grasp, but exhibiting a preference does not mean you are biased in any other area save that preference. You want to infer that because a newspaper takes a certain editorial page position, that everything printed in that paper is affected/tainted by that position. Not only does that imply that anyone who takes a position with which you disagree is incapable of independent thought on any other issue, it ignores the fact that any endeavor like the Times is made up of many human beings, all of whom have their own thoughts based on their life experiences, upbringing and education.(Do you recall Judith Miller, for example? Have you read Thomas Friedman’s op-ed pieces on the Iraq war? Do you actually read the paper at all, ever?) Disagreeing with you is not in itself prima facie evidence of bias. Perhaps your comment was made while gazing in the mirror?

        (For the record, I voted differently than the NYT endorsement in 1980, 1996, 2000, and 2012, though in no instance did I vote for a Republican.)

      • Marty - Nov 1, 2012 at 6:18 PM

        I’d say exhibiting preference is the very definition of bias, but hey I’m not trying to convolute like you are.

        You have a point about editorial bleeding over to the news pages, but your first statement is so fucking dishonest I don’t care to waste time considering anything else you write.

      • churchoftheperpetuallyoutraged - Nov 1, 2012 at 6:38 PM

        Perhaps the NYT isn’t as biased as Fox or MSNBC, but they haven’t endorsed a Republican for president during my lifetime.

        Who cares when the last time the NYT endorsed a Rep candidate, answer the damn question. When’s the last time that the NYT, or another major [liberal] media source blantantly lied about a story or intentionally misled viewers?

        For reference, here’s a few things that Fox News has done:

        Fox News wonders why people haven’t moved on from the Phone Hacking scandal in Britain*. The gentleman they bring on says “News of the World is a hacking scandal….why haven’t we moved on? Citicorp got hacked, insurance companies got hacked into….we have to figure out a way to deal with this hacking problem!”

        How about, YOU WERE THE ONES DOING THE HACKING! while everyone else was a victim of it.

        Or how about accusing the gov’t of fudging the jobs report from last month to make Obama look better? Or the lies about Benghazi? Or the stupidity of Obama’s birth certificate?


      • Marty - Nov 1, 2012 at 6:59 PM

        Church, I’ll expound on this when I have more time, but MSNBC has willfully edited Romney speeches and the Treyvon Martin 911 calls to conform to their agendas. As for the NYT, I don’t read that paper for news stories so I can’t mention specifics beyond their columnists am editorials.

        And sorry we disagree about Benghazi, but there is certainly a coverup. Even Bob Woodward is investigating. How you’d lump that in with your other slightly valid quips makes me think you have some serious levity issues.

      • Marty - Nov 1, 2012 at 7:00 PM

        …ad by levity I mean you’re a fucking hack.

      • churchoftheperpetuallyoutraged - Nov 1, 2012 at 7:42 PM

        And sorry we disagree about Benghazi, but there is certainly a coverup

        A coverup by the presidency or a coverup by the media? Because we aren’t talking about the former, we’re talking about the latter. Point taken on MSNBC and Trayvon Martin.

        And how am I a hack?

      • philsieg - Nov 1, 2012 at 9:14 PM


        Is your understanding of the written word so poor that you fail to grasp all the meanings of what I wrote? Of course one can exhibit a preference in one area and that preference can lead to a bias in all areas. You are a perfect example.

        Conversely, one can state a preference in one area (i.e.for president) and still maintain objectivity in other areas. It’s called an open mind, but from your response I doubt you know much about that.

        To phrase it another way, to exhibit a preference for something is to be biased in favor of that something – but only that something. To say that a single preference leads to bias in all areas assumes facts not in evidence. The fact that you see my statement is dishonest reflects poorly on you and your education – not me.

      • Marty - Nov 2, 2012 at 12:29 AM

        Philsieg, try something more original than patting yourself on the back for being open minded. Just because because you’re a committed Democrat doesn’t mean you have to act like a self righteous dumbshit. I have seen nothing here to hint at you considering anything other than sniffing your own ass and approving.

      • philsieg - Nov 2, 2012 at 3:20 AM


        So it is the reading comprehension thing, after all. It’s pretty clear from what I’ve written here that I’m not a Democrat.

    • nctaxpro - Nov 1, 2012 at 4:09 PM

      Rasmussen is well known for leaning toward the Republicans, just as PPP leans toward the Democrats.

    • dprat - Nov 2, 2012 at 1:37 AM

      Fiorentino is every bit as clear-headed, rational, evidence-based and unbiased about Rasmussen Reports and “a blogger on” as he is about Ryan Howard. It’s that kind of consistency that both I and Hall-of-Famer Joe Morgan admire so much.

  3. thefacts121 - Nov 1, 2012 at 3:29 PM

    I don’t think this accurately characterizes much of the criticism of Silver’s modeling. The basic critique that I’ve seen is that Silver relies too much on state polling, which often features smaller samples and is done by less reliable pollsters, and conversely under-relies on national polling featuring larger samples from more historically accurate pollsters. Silver has his reasons for doing what he’s doing (in certain years the state polls have been more accurate) and the critics have points as well (that these are not the same state pollsters who were accurate in past years and that many of the current state pollsters appear to have skewed samples with regard to partisan identification). In any event, I think there are two key conclusions to draw: 1) you can make anyone look good by ignoring real criticism and only looking at their silliest critics; and 2) just because there analysis conducted under the rubric of math does not make any criticism of it a desire to return to gut feelings.

    • bravojawja - Nov 1, 2012 at 3:45 PM

      The reason Silver uses state polling rather than national polling is because he’s trying to figure out who has the best chance to become President. What matters is who wins which state and therefore the Electoral College. I’m sure President Al Gore can tell you what winning the national popular vote means.

      • thefacts121 - Nov 1, 2012 at 4:07 PM

        He actually uses state polling as a large part of calculating the total national vote projection. In other words, President Obama is winning in most swing state polls and losing in most of the national polling. Silver projects that Obama will win the popular vote in addition to the electoral college based on superior state polling results despite the inferior national polling. He thinks the odds of an electoral college/popular vote split are pretty small (less than 8% I believe in his latest model).

        The point is, as Silver has said himself, the national polling (advantage Romney) reflects a different state of the race than the state polling (advantage Obama) and we’ll know in a few days which was more accurate. I don’t think it’s a Luddite position to say that national polling has some advantages on state polling, which is where most of the non-straw man criticism comes from.

      • philsieg - Nov 1, 2012 at 4:24 PM

        “I don’t think it’s a Luddite position to say that national polling has some advantages on state polling, which is where most of the non-straw man criticism comes from.”

        What would those advantages be in a system that decides the winner based on the electoral votes associated with states won? Only two states – Maine and Nebraska – can split the electoral vote between candidates, and that depends on carrying congressional districts.

        (BTW, Nate has the likelihood of an electoral/popular vote split down to 5% currently. The president is now projected to have a 79% chance of winning the election with 50.5% of the popular vote and 300 electoral votes. Michael Bloomberg and Chris Christie certainly haven’t helped the Romney narrative.)

    • Jeremy Fox - Nov 1, 2012 at 4:06 PM

      Individual state polls have smaller sample sizes than national polls, but *collectively* the sample size is much larger.

      The broader point is that cherry-picking information is a terrible idea. Just as major league teams use stats and scouts to advise them on who to draft, sign, and trade, good election forecasters draw on lots of polling information, not just national polls or the polls from one particular pollster.

      • thefacts121 - Nov 1, 2012 at 4:48 PM

        No argument from me on your cherry picking point. I think the legitimate criticism of Silver’s model comes from an argument that he is incorrectly weighing the relevant data points, not that he includes these different data points at all.

        If someone presented a version of WAR that was otherwise like Fangraphs model but quintupled its current weighting of baserunning value, it would not be a criticism of math itself to say that this model was flawed if the justification behind the weighing seemed faulty.

      • Jeremy Fox - Nov 1, 2012 at 5:56 PM

        Well, except the people criticizing him for incorrect weighting aren’t doing so on the basis of evidence. They just don’t like the conclusions that Silver’s weighting leads to. Silver has data-based reasons, which he’s described, for weighting the state polls as he does. Just as there are data-based reasons for, say, the weight that baserunning is assigned in fWAR. Not there isn’t scope for reasonable disagreement about weightings. Indeed, there are other quantitative forecasting sites, like Votamatic, that weight state-level polls even *more* heavily than Silver does. But I have yet to see anyone make a *data-based case* that Silver’s weightings of state-level polls (not to mention other sites’ even higher weightings) are unreasonable, or that there is scope for *reasonable* weightings that would give Romney a better-than-even chance of victory.

  4. fusionix7 - Nov 1, 2012 at 3:31 PM

    Thanks for bringing him up. I saw an interview with him last night, very bright guy, and a Tigers fan. I forgot about him until now. His book should be an interesting read.

  5. thebadguyswon - Nov 1, 2012 at 3:33 PM

    I know it’s the slow time of year, but do you really have to delve into anything politics-related?

    • Craig Calcaterra - Nov 1, 2012 at 3:50 PM

      The worst in sports and sports writing come from people who act as if nothing else exists in the world except sports and sports writing.

      • thebadguyswon - Nov 1, 2012 at 5:42 PM

        Yeah right. Or maybe people want baseball on their baseball sites. Don’t assume badguys doesn’t follow politics.

      • thebadguyswon - Nov 1, 2012 at 5:44 PM

        Admit it Craig, you have nothing to write about.

      • Craig Calcaterra - Nov 1, 2012 at 5:50 PM

        I’ve written 18 posts today. Aaron has written 13 or 14 more. Exactly one of those posts was about something other than baseball, and this one focused on the work of a person who has been extremely influential in baseball and covers a subject (inside-outside reporting) which is of great relevance to baseball reporting. And which is something I have written about since the day this blog began.

        But sure, I got nothing to write about.

      • thebadguyswon - Nov 1, 2012 at 8:44 PM

        The problem with a thread like this, in my humble opinion, is then you open it up for people to divert from baseball and blow smoke up everyone’s ass about their political beliefs. I mean what’s next – baseball and the Catholic priest scandals? I go to Real Clear Politics more than I visit here. That’s where I want politics. Here I want baseball mixed with the usual humor and snark. That’s the only point I’m trying to make.

  6. Marty - Nov 1, 2012 at 3:41 PM

    I’ve seen Silver criticized from the middle left as well. Gil Gross comes to mind. Technologists often become committed to ther methods, blind to flaws. He may be more right than wrong. But building this firewall around him to protect his analyses from being scrutinized is likely a result of liking his conclusions.

    • philsieg - Nov 1, 2012 at 4:12 PM

      FWIW, my defense of Nate comes from the fact that I’m a scientist with a background in mathematics and the life sciences who has had to develop predictive models as part of my professional career. I respect his diligence and his results.

      As far as liking his conclusions goes, I voted for Jill Stein because Obama’s human rights record is deplorable and neither candidate has even mentioned the most pressing issue facing the nation and planet for the duration of this campaign.

    • Jeremy Fox - Nov 1, 2012 at 4:19 PM

      Um, nobody’s building a firewall around anybody. Go ahead and criticize him (and I say that as someone who share’s his politics). But don’t criticize him because *you* don’t like his conclusions, or don’t understand how he arrived at them and assume that anything you don’t understand must be rubbish.

      Different quantitative election forecasters actually do disagree with Silver in various ways, for instance on the issue of whether one should attempt to estimate and correct for things like pollster house effects, or whether it’s better to just ignore complications like house effects and hope that you’ve got enough data that the signal emerges. But as Craig correctly notes, those sorts of technical statistical issues emphatically are *not* the issues that the vast majority of Silver’s critics have been raising.

    • churchoftheperpetuallyoutraged - Nov 1, 2012 at 4:21 PM

      He changes his model all the time, so he’s not just wed to over iteration. Also, can we ginger him the benefit of the doubt at least? 49/50 states currently predicted in 2008 and predicting huge gains the GOP in ’10 aren’t enough?

      • churchoftheperpetuallyoutraged - Nov 1, 2012 at 4:26 PM

        Damn phone, should be one iteration and give him, not ginger

        Edit function!

      • Marty - Nov 1, 2012 at 4:59 PM

        Were his 2010 predictions that much more weighted in favor of the GOP than other analysts at the time? I think that’s partly why he comes under scrutiny.

        Perhaps he’s boy wonder with the secret formula. I doubt even few are qualified to say either way. That’s my point. I think the weight people put on his analysis is influenced by who you have in the race. This has always been the case since polling and politics became joined at the hip.

        But just to be clear, I’m not saying you have to be pro Obama to agree with Silver’s analysis. The GOP were probably all boners all the time over his 2010 predictions.

      • Jeremy Fox - Nov 1, 2012 at 6:15 PM

        No Marty, the weight you put on Silver’s analysis is *not* just a function of who you have in the race. I share Silver’s politics–and I bought his forecasts in 2010 too, even though doing so was depressing. And I was far from alone in this. Democrats didn’t noisily criticize Silver in 2010 the way Republicans are in 2012.

        There really is a difference between the two sides here. Lashing out at quantitative, evidence-based analyses because you don’t like the conclusions really is more of a Republican sport, at least when the analyses concern election outcomes.

      • Marty - Nov 2, 2012 at 12:35 AM

        I suppose mistaking criticism for “lashing out” is a defect of the insecure, irrespective of party affiliation. Can we have disagreements without broad brush straw manning, eh Democrat?

  7. cur68 - Nov 1, 2012 at 3:45 PM

    I love your politics. Election Time USA is better than any reality show. Incidentally I saw Mr. Silver on your news program, The Daily Show, and he struck me as intelligent, insightful, and open about his processes but aside from that I have no idea if he’s any better or worse than any other poll pundit. I look forward very much to your election itself and finding out who the winning tribe…er…party will be and whom will be your Chief…er…President. Must go: sled dogs need feeding….

    • stex52 - Nov 1, 2012 at 3:54 PM

      Cur, we’re buddies, right? You know I have really enjoyed the times I visited Canada, Right? So I can say this.

      Whenever you talk American politics, I think of a magazine insert that ran in the National Lampoon circa 1973 (I was in college).

      “Canada, the Retarded Giant on Your Doorstep”

      Love you guys. But it was very funny satire, too.

      • cur68 - Nov 2, 2012 at 1:54 AM

        Hey! I resemble that remark…

    • indaburg - Nov 1, 2012 at 4:24 PM

      The Daily Show is my only trusted source for information. I also saw that Silver interview and he did seem to know his stuff.

  8. dowhatifeellike - Nov 1, 2012 at 4:05 PM

    Political pundits don’t like the numbers Silver is pumping out because he is destroying their “close race” narrative. Political news attracts a lot more eyeballs when they pretend it’s anybody’s race.

  9. churchoftheperpetuallyoutraged - Nov 1, 2012 at 4:17 PM

    Here’s another article that parallels what Silver, Cameron @ FanGraphs, and Schatz @ have faced in their earlydays

  10. Jeremy Fox - Nov 1, 2012 at 4:28 PM

    Just wanted to say great post Craig. I’d noticed the parallel too, I think it comes up in a lot of areas of life. Even within many scientific and social science fields. Speaking from personal experience as a scientist, lots of fields of science are becoming more quantitative than they used to be, which annoys older scientists who aren’t as numerate as the next generation. This is a great example of why I read HBT–you’re aware of and think about stuff besides just baseball.

  11. lazlosother - Nov 1, 2012 at 5:18 PM

    So, the Saber guys like Obama because he has a higher WAR, and the inside baseball guys like Romney because of intangibles?

    These interwebs ideas are far too confusing. I’m voting for Mike Trout.

    • churchoftheperpetuallyoutraged - Nov 1, 2012 at 5:27 PM

      I voted for Kodos

    • thinman61 - Nov 1, 2012 at 8:00 PM

      I’m pretty sure this is the first (and last) time anyone has even indirectly categorized Mitt Romney as a dirt dog.

  12. crackersnap - Nov 1, 2012 at 5:50 PM

    The problem, as I see it, is that those active in the political process cannot afford to have a momentum of the narrative begin to build against them. It doesn’t matter which side of the fence you are on when Silver’s numbers come out against you. You are forced to reject those conclusions lest they create any synergistic tide of self-fulfillment.

    In fact, the greater the proclamations are against you, by anybody, the stronger your side needs to bark out against that information in order to create backlash, hopefully to create a counter tide of narrative in your favor prior to the actual deadline of voting day.

    • dowhatifeellike - Nov 1, 2012 at 5:59 PM

      That is the way they play the game… I just wish people would stop buying what they’re selling. The practice will end only when elections become more about substance than style.

  13. klokskap - Nov 1, 2012 at 10:23 PM

    Nate Silver is a really smart guy. That doesn’t make him right. There are other poll analysts out there who come to much different conclusions than Silver:

    • klokskap - Nov 2, 2012 at 11:38 AM

      Baseball Crank is spot on of his analysis of using polling top line data v. internals, comparing the polling analysis to PECOTA and other events where people misread the underlying data – especially as told in Michael Lewis’s great books “The Big Short” and “Boomerang: Travels in the New Third World”.

    • djpostl - Nov 7, 2012 at 3:43 AM

      And those other guys were dead wrong…

  14. djpostl - Nov 1, 2012 at 11:01 PM

    The idiot Tea Party trolls hit Twitter. And attacked the wrong guy named Silver. Priceless.

  15. djpostl - Nov 7, 2012 at 3:41 AM

    For all the shit he caught…dude was 50 for 50 state-by-state (he went 49 for 50 in 2008)

Leave Comment

You must be logged in to leave a comment. Not a member? Register now!

Top 10 MLB Player Searches
  1. A. Rodriguez (4106)
  2. J. Cueto (3709)
  3. C. Hamels (3599)
  4. T. Tulowitzki (3127)
  5. M. Trout (3103)
  1. S. Kazmir (3057)
  2. M. Conforto (2975)
  3. D. Price (2968)
  4. C. Gonzalez (2942)
  5. J. Reyes (2938)