Friday, December 21, 2012

What's So Special About $13 Million?

I should start off by saying that if this feels a little conspiracy-theory-ish, that's because it is. I may be being too paranoid, but no one else is really talking about this, so I thought I would at least point it out.

Dedicated observers of the hot stove have been experiencing déjà vu this MLB offseason. If you've been feeling unlucky, it's not just in your head; the number "13" has popped up unusually frequently in final contract numbers:

Player New Team Dollars Length AAV
David Ortiz Red Sox $26 million 2 years $13 million
Torii Hunter Tigers $26 million 2 years $13 million
Dan Haren Nationals $13 million 1 year $13 million
Mike Napoli Red Sox $39 million 3 years $13 million
Shane Victorino Red Sox $39 million 3 years $13 million
Ryan Dempster Red Sox $26.5 million 2 years $13.25 million
Edwin Jackson Cubs $52 million 4 years $13 million

This doesn't count Andy Pettitte (signed with the Yankees for one year, $12 million with incentives that could bring him to $13 million), Kevin Youkilis (signed with the Yankees for one year, $12 million), and RA Dickey (signed an extension with the Blue Jays for two years, $25 million).

What is going on here? At first, I figured it was a coincidence, but now I'm not so sure. There certainly seems to be an informal cap on the annual salaries given out to free agents this year. Oh, to be sure, elite free agents have broken the bank, such as Zack Greinke's $24.5 million AAV and Josh Hamilton's $25 million. Teams rightly have no problem breaking the magic barrier and doing everything in their power to secure a superstar—but the mid-level free agents have a very specific value on the free market, it seems.

This raises the question of whether it really is a free market. To be clear, I am not making an accusation—but any time the numbers line up like this, it raises the specter of collusion. Collusion may seem extreme and far-fetched, but it has been pretty common throughout baseball history, including recently, if you believe the MLBPA. It's actually not that outrageous of a possibility. What we know is that there has been an unusually high number of identically valued contracts this offseason, whether by secret, explicit arrangement between teams or an unspoken consensus around the league that $13 million is where the bidding stops.

In the mystery of whether there is anything deeper—or sinister—behind this study in numerology, a potential clue is the revamped system of free-agent compensation in the new CBA. (If you're not familiar with it, a good explanation is here.) The value of the "qualifying offers" that teams extend to their free agents under this system is calculated from the average of the 125 highest player salaries. In another eyebrow-raising coincidence, the value of a qualifying offer this year was $13.3 million.

There are numerous inferences to draw here. The simplest is that teams are simply trying to artificially depress the value of the qualifying offer (or at least keep it steady at roughly $13 million). That could be part of it, but there are also more complex forces potentially at play here.

To date, only five players (Josh Hamilton, Zack Greinke, BJ Upton, Aníbal Sánchez, and Hiroki Kuroda) have signed for higher AAVs than the value of a qualifying offer. Others (Nick Swisher, Michael Bourn, Kyle Lohse) figure to do so as well in the near future. All of these players were either extended qualifying offers or were not eligible for them (due to being traded midseason). In contrast, no player who didn't receive a qualifying offer is expected to get a higher AAV than $13 million.

Indeed, that's exactly what they have been getting: $13 million per year. That's a sign that maybe the market for non-qualifying-offer players is still strong—perhaps strong enough to reach $14 million or $15 million if unencumbered. But teams have an incentive to encumber—and to set the "ceiling" for these B-level free agents' salaries at a number just a tad below the value of a qualifying offer.

The incentive is to discourage other teams from making qualifying offers in the future. If any non-qualifying-offer free agent did receive a contract bigger than $13.3 million, teams would take note of this missed opportunity to gain a draft pick and might be more liberal in extending qualifying offers after 2013 or 2014. A given team doesn't want the other 29 to realize and do that, however, because it means that it would have to give up a draft pick if it wanted to sign that free agent. The fewer qualifying offers that are extended, the more aggressive teams can be on the free-agent market because the fewer draft picks they'll lose in doing so.

This scenario assumes that the clump of contracts around $13 million is meant to influence other teams' decisions on extending qualifying offers. But it could also be a way to influence players' decisions on whether to accept them.

Say MLB teams are all colluding to keep non-qualifying-offer free agents at AAVs under $13.3 million. The flip side of that is that teams are allowed to go berserk over qualifying-offer free agents; they're the only ones left to throw money at. This ensures that no qualifying-offer free agent ever settles for an AAV less than $13.3 million. Then, in future years, players who receive qualifying offers look at the history of past free agents in the same position and see that they would be stupid to accept one year at "only" $13.3 million. It then becomes easier for teams to extend qualifying offers—and thus easier to secure an extra draft pick—with less of a fear that their players will accept the offer, which the team may not truly be interested in paying.

This is a rather opposite scenario from the other; both seem like plausible possibilities, though. I'm sure others, even more complicated, could be thought up too. I don't presume to know the real reason for the cluster at $13 million, and, again, I'm not making a specific accusation. It's worth some very critical thought, though.

Monday, November 19, 2012

Crystal-Ball Report Card: 2012 Season

We're entering a lull in prediction season, but that doesn't mean we're done talking about predictions. It's easy for people to make forecasts, but disappointingly few follow up on them to see how they shook out—and, even then, it's usually only the people who hit the nail on the head looking to gloat. In the name of accountability, I wanted to look back on the predictions I've made over the past year—and hopefully learn something about more accurate forecasting in the future as a result.

Specifically, I want to take a look at my predictions for the 2012 MLB season, broken down by division: AL East, AL Central, AL West, NL East, NL Central, and NL West. More recently, I also made picks about how I thought the November elections would go, but (a) I've hashed through those pretty well on Twitter and (b) I don't think an article in which I gloat about hitting the nail on the head would be very interesting. My MLB predictions, on the other hand, were much more of a mixed bag.

Back in March and April, I calculated the specific win-loss records that I thought each team would end the season with. Let's start by looking at how I did with the raw numbers:


Next we'll dive into some of the specific claims I made.

Prediction: Both the Orioles and Athletics would stink up the joint on their ways to respective last-place finishes.
What Really Happened: In my defense, I said that the Orioles were "hardly an atrocious team and have the potential to play some watchable ball this year"; I called the Athletics "the division's most interesting—and most unpredictable—team." But no one could have predicted all those Ws. The A's and O's stunned the baseball world by threatening all season to win their respective divisions—two of the strongest in baseball. My error was in overestimating the strength of some of the other teams in those brackets. I was convinced that the Rangers and Angels would both be powerhouses; instead, the $55 million A's stole the division crown from both of them. I was particularly wrong about the Angels, whose shaky starting pitching caused them to win 11 fewer games than I thought they would.

Prediction: The Red Sox would not be a playoff team, but they would win 88 games.
What Really Happened: The Red Sox lost 90 games for the first time since 1965. This was a team that completely imploded under the leadership—if you can call it that—of Bobby Valentine. More relevantly, Boston's pitching just fell apart. I was skeptical about the team because I saw that it had "only two sure-thing starters (Jon Lester and Josh Beckett)" plus a handful of pitchers with promise. It turns out that none of that promise was fulfilled (Daniel Bard, in particular, was a mess) and their sure things forgot how to make outs. At the very least, my observation that, "when they bled, they could not clot" proved accurate, as the franchise hemorrhaged losses and fans all year long.

Prediction: The Tigers would win the AL Central, but more by default than domination; the team would actually be kind of mediocre, with only 89 wins.
What Really Happened: Exactly that... sort of. The Tigers limped to the division crown with 88 wins thanks to the same liabilities that I predicted: a ghastly defense and a middling offense. I correctly called Alex Avila's and Jhonny Peralta's falls back down to earth, though I also wrongly called out Austin Jackson for being an underachiever. But then, of course, the whole team made me feel a little silly when it dominated its way to the World Series.

Prediction: The White Sox would finish second in the AL Central. Jake Peavy, Chris Sale, Gavin Floyd, and John Danks would form a dominating starting rotation; Adam Dunn would bounce back to become a middle-of-the-order threat.
What Really Happened: The White Sox did even better than I expected (I still had them finishing slightly below .500), getting bounceback seasons out of not only Adam Dunn, but also Alex Ríos, whom I had given up on. Gavin Floyd wasn't reliable, and John Danks was lost to injury, but the rest of the pitching staff stepped up to fall just short of the fewest runs allowed in the division. I was particularly prophetic on the seasons of prospect Sale and injury question mark Peavy.

Prediction: Of the top four teams in the NL East, "each team is capable of dominating to the tune of 100 wins (yes, even the Nationals), and each team could collapse like a house of cards to below .500 (yes, even the Phillies). The one certainty—and, in my opinion, the safest bet in all of baseball this year—is that a certain team from New York will sink comfortably to the bottom."
What Really Happened: What was certain was wrong, and what seemed fanciful became reality. The Mets finished fourth in the NL East, guaranteeing my premature retirement from sports gambling; I foresaw Johan Santana's inconsistency and Jason Bay's suckitude but failed to account for David Wright's resurgence and the force of nature that is RA Dickey. Meanwhile, the Nats came close to those 100 wins and the Phillies finished at exactly .500. This fascinating division deserves some more broken-down analysis, though...

Prediction: Reports of the Phillies' demise would be greatly exaggerated. While the loss of Roy Oswalt would take a few wins off their 2011 total, they would still be a dominating pitching team and the class of the National League.
What Really Happened: Uh, oops. Many others saw this coming, but I guess I missed the warning signs. Roy Halladay lost time due to injury, and even when he pitched, he was mediocre (4.49 ERA); meanwhile, Cliff Lee forgot how to win (only six of them over a full year). As a result, they had only the division's third-best pitching. I was slightly redeemed when it turned out that the Philadelphia offense, which everyone else saw as ripe for a collapse, only weakened incrementally—consistent with what happens when a team gets one year older. Still, my claim that Philly would have "the best offense in the division" due to others' weakness was way off base (they had the third-best).

Prediction: The Nats would be the breakout team of 2012, pitching their way to a playoff berth. The additions of Edwin Jackson, Gio González, and Stephen Strasburg (back from injury) would add anywhere from nine to 18 wins to their total of 80 from 2011, and they would possibly lead the majors in ERA.
What Really Happened: It was 18 wins, not nine, and they had the second-best ERA in the majors, but I nailed everything else. The Nats as contenders was one of the preseason predictions I argued most forcefully for, and I didn't see a way (beyond injury) that their fearsome foursome of starters wouldn't improve the Nats' fortunes dramatically. What I didn't see was that their offense would mature as well; they had the division's best. Maybe I should have, though—I specifically predicted that Adam LaRoche's 25-home-run power would return (he hit 33).

Prediction: Brandon Beachy and Mike Minor—not the injury-prone Tim Hudson and Tommy Hanson—would lead the Braves to a Wild Card berth.
What Really Happened: Beachy was dominant but was lost to Tommy John surgery in June. Minor ended up struggling to a 4.12 ERA, and Hanson was actually healthy all year—though he also provided lukewarm results. It was pitchers who came out of seemingly nowhere to give the Braves their boost: Paul Maholm, acquired in a trade from the Cubs, and Kris Medlen, whose 1.57 ERA in 138 innings remains the seminal stat of the 2012 Braves season. But hey, at least I got the Wild Card berth right.

Prediction: The Marlins would finish in fourth place with only 85 wins, the victims of overrated offseason signings.
What Really Happened: The Marlins did a lot worse, losing 16 more games than I thought. But, in my defense, the tone of my Marlins preseason rundown was hardly positive. I saw the José Reyes signing as akin to adding an average shortstop, the Heath Bell signing as basically pointless (most relievers are), and the Mark Buehrle signing as useful, but only to replace the likely-to-be-injured Josh Johnson (for a net gain of zero). Reyes ended up avoiding the DL and being a quality regular, but he couldn't make up for the absence (in spirit and then in reality) of Hanley Ramírez's potent bat from the middle of the lineup. Where I was most wrong was in saying that, while they would not make the playoffs because they had so much ground to make up, "this is an improved team, no question"; turns out there was a question, as they lost three more games than in 2011.

Prediction: The Brewers would win the NL Central. Aramis Ramírez would replace Prince Fielder's pop in the lineup, and their 2011-division-winning pitching staff would continue to be quietly solid.
What Really Happened: I truly believe the Brewers would have won the division again in 2012 if their pitching staff—specifically, their bullpen—hadn't been so very loudly awful. The Brewers actually scored 55 more runs in 2012 than they did with Fielder in 2011; Ramírez filled in very nicely, turning in an even better season than his excellent 2011. Their starters also had a 3.99 ERA, again jibing with my prognostication. Their bullpen, however, was the league's worst, with a 4.66 ERA, 33 losses, and 29 blown saves! With bullpen performance being one of the most fickle things about baseball from year to year, if Milwaukee could play the season over, I think they'd have a different result. Look, too, for them to improve almost automatically in 2013.

Prediction: The Reds were overhyped and would finish third with 86 wins; 90 would be their ceiling due to a mediocre starting staff.
What Really Happened: The Reds made me look like an idiot; Cincinnati was this close from finishing with the top record in baseball. What did me in was the rock-solidness of the Reds rotation; their main five started 161 of the team's 162 games. Bronson Arroyo and Homer Bailey both spun phenomenal seasons, especially considering the ballpark they call home, and Mat Latos was much more of an impact player than I had foreseen. I never would have dreamed that together they would give up the fewest runs in the National League.

Prediction: This would be the year the Pirates seriously challenge for a winning season, but ultimately they would fall short, with 76 wins.
What Really Happened: The work of the devil, apparently. The Pirates stormed off to a great start, realizing more potential than even I thought they had in them. But, famously, the Bucs stopped there, plummeting to a 79-win finish. While I couldn't have predicted the wild ride, my final guess was pretty well on target. My February argument about the strength of the Pittsburgh rotation also proved prescient as the reason for Pirates fans' early-summer hope. It turns out they only had the NL Central's third-best pitching, however, not second-best as I had envisioned.

Prediction: The Astros, Cubs, and Twins would be among the worst teams in baseball, winning 55, 62, and 67 games respectively.
What Really Happened: The Astros, Cubs, and Twins were among the worst teams in baseball, winning 55, 61, and 66 games respectively. Sometimes, the worst teams are the easiest to predict, though I did put my neck out a bit when I said only four Twins would hit double-digit home runs (Joe Mauer, Justin Morneau, Josh Willingham, and Danny Valencia). I was wrong about Valencia and failed to include Trevor Plouffe (whose sudden power surge must qualify as the surprise of the 2012 AL Central) and Ryan Doumit (who barely made it himself, hitting 10 homers), but the basic idea held true: Target Field saps power.

Prediction: "I don't see how [the Diamondbacks' excellent 2011] couldn't be [for real], though, for it was built on an extremely solid foundation... it would be more surprising if [Ian Kennedy and Daniel Hudson] regressed this year, considering the promise that was held for them when they were minor leaguers." The Diamondbacks would once again win the NL West, though with only 88 wins in this weak division.
What Really Happened: The Diamondbacks did return the NL West's second-best offense and second-best pitching. That should have been good for second, if not first, place, and indeed it did produce a Pythagorean record of 86–76. But reality intervened, and the DBacks finished third with 81 wins—not an altogether terrible prediction, but clearly a miss. To blame were Daniel Hudson's Tommy John surgery, Justin Upton's average output, and Ian Kennedy's blah 4.02 ERA.

Prediction: The Rockies would be MLB's biggest surprise in 2012. The rotation would put it all together for at least 85 wins. "If things break right, Colorado could run away with the division title."
What Really Happened: Things, ah, did not break right. The Rockies lost 98 games and were my worst overestimation of the offseason. Specifically, that starting staff I was so optimistic about was so bad that the team switched to a four-man rotation, limiting starters to 75 pitches each. It was a failed experiment, and Colorado starters finished with a 5.81 ERA. Drew Pomeranz did not dominate over a full season as I boldly predicted, and my diamonds in the rough Jeremy Guthrie and Jamie Moyer lasted a collective five months in the rotation. Finally, Jhoulys Chacín and Juan Nicasio both failed to come back from injuries, leaving the pitching cupboard bare. The best laid plans of mice and men...

Prediction: The Dodgers would fight to stay out of the cellar, lacking any kind of supporting cast for Matt Kemp and Clayton Kershaw. They would finish with 74 wins.
What Really Happened: The Dodgers loaded up on cash and bought (literally) a whole team, including Hanley Ramírez, Shane Victorino, Adrian González, and Joe Blanton. I was actually right about James Loney and Dee Gordon crashing and burning this year, but I didn't expect all-stars to take over their positions. Personally, I don't think anyone can be held to their preseason Dodgers prediction, since the team that ended 2012 in Chavez Ravine simply bore no resemblance to the one that started it. A solid rotation was also something I underestimated, however, as the Dodgers played the role of Cincinnati with four consistent starters with ERAs under 3.73.

Prediction: The Giants "appear to have hit a ceiling with their two-way low-score strategy." Without a reliable offense, they would limp to 84 wins and third place. Melky Cabrera would return to being an out machine, though Aubrey Huff would manage to resurrect his career (yet again). Ryan Vogelsong would discover mediocrity, and Barry Zito would continue his.
What Really Happened: The 2012 World Series champions, that's what happened. Virtually all my predictions turned out the exact opposite: Zito found new life, Vogelsong seems to have achieved a new normal, Cabrera was unreal (as was, it turns out, his newfound muscle), and Huff turned in fewer than 100 at-bats. Meanwhile, Buster Posey led a truly shockingly good offense of the kind San Francisco has lacked since Barry Bonds. I could've told you in the spring that, with that kind of offense, the Giants would be World Series favorites. But I couldn't and I didn't—and that's why you can't predict baseball.

Sunday, November 4, 2012

Predicting the 2012 Election

In a recent post, I explained why "calling" races and making predictions based on qualitative, not quantitative, factors was iffy at best and academically irresponsible at worst. So, naturally, what follows will be my arbitrary and binary predictions for the 2012 elections.

(In seriousness, I do want to make it clear that this is in no way a scientific prediction. But I also don't think there's anything wrong with that, as long as it's acknowledged up front and everything that follows is kept in its proper perspective. Making predictions, as pundits do on Baseball Tonight or the World Series pregame show or whatever, is fun, and it's nothing I would begrudge anyone. I just wish punditry would be seen for what it is—entertainment.)

House


Here, my prediction will be the roughest and, probably, the most uninformed. Unfortunately, I haven't had a chance this cycle to look at the competitive House races in much depth. In 2006, 2008, and 2010, we had the benefit of knowing it was a "wave election," but there is no such trend this year. The consensus among experts is that there may be a slight Democratic edge (as in the Senate and presidential races), but it's close to a draw. If the parties split the tossup seats, Democrats would net only a handful—zero to five seems to be the generally agreed-upon range. Looking at the roster of competitive races, though, I find a bit more to like for Democrats. Going through this list, keeping score based on my sense of each race, and then splitting my own personal "tossups" 50-50, I make a back-of-the-napkin guess that the next House will consist of 202 Democrats and 233 Republicans.

Gubernatorial


In the underrated governors' races, I have already provided my overall rankings in the form of a spectrum accessible via the tab on the top of the page. This year, there are four clear Democratic favorites and four clear Republican favorites; you can see my picks for those on the Gubernatorial Rankings page. But what of the three tossups: Washington, New Hampshire, and Montana?

Well, the way the chart works is that I order the races from most Democratic to most Republican, so the tossup closest to the blue side is the one most likely to go blue and vice versa. However, I should emphasize that the reason I ever place any race in the "tossup" category is that it truly is too close to call—by definition, that there's nothing that gives me a hint the race is leading one way or the other. I'll pick them for you here—again, for entertainment purposes only—but know that it's a stab in the dark.

I'm going to say that Montana's next governor will be Republican Rick Hill, while Washington and New Hampshire will elect Democrats. In Montana, I expect that presidential coattails (i.e., the fact that Mitt Romney will win the state handily) will outweigh the coattails of an outgoing Democratic governor who's not on the ballot. In Washington, likewise, Barack Obama's easy win should help drive Jay Inslee–inclined supporters to the voting booth. Furthermore, Inslee has enjoyed a slight polling lead for a few months now, with the race only just recently drifting into the margin of error. In New Hampshire, the opposite has occurred, with Democrat Maggie Hassan coming off a strong set of polls showing her up four points, five points, and five points after a tied race for much of the duration. My prediction below (spoiler alert!) that Obama will win the Granite State can also only help her. Here's what I expect the gubernatorial map to look like at the end of election night:


Senate


In the Senate, I also have a useful chart explaining which parties I expect to win which races, and how safe that prediction is. Obviously, I won't be picking against myself in any of those races that I've categorized (otherwise I'd just change the chart). But the chart does leave five races as tossups, and, yet again, I will make some random predictions for you.

My crystal ball will again be generous to Democrats, as I perceive the median voter's mood to be ever-so-slightly left-leaning at the moment, so I give them the top four tossups in the chart. In Massachusetts, I've thought that Democrat Elizabeth Warren would prevail from day one; Republican Senator Scott Brown eked out a win in 2010 only because turnout was so low and Republicans were so motivated to vote in that special election. The 2012 voter-turnout model, of course, will look very different in this deep-blue state. Still, Brown has proven to be probably the strongest Senate candidate in the entire country, keeping the race tight all year. It has only been recently that Warren has pulled ahead—slightly—in polling.

Virginia and Wisconsin both figure to feature razor-thin margins as well, and I'm not convinced that either candidate has an advantage as we stand today. On Election Day, though, it'll come down to turnout, and I expect Democrats to have the superior ground game in both states on Tuesday. This is in large part because both states are important swing states in the presidential, and (spoiler alert!) I view Obama as the favorite in both. That should be enough to pull Tim Kaine and Tammy Baldwin into the Senate.

A very interesting case is Indiana, which, along with Missouri, looks as though it could be one of two states where the Tea Party costs Republicans a Senate victory—2010 redux. This went from a safe Republican seat when it was Richard Lugar's to lose, to a lean Republican seat when the very conservative Richard Mourdock became the nominee, to a complete tossup (some would say leans Democrat!) when Mourdock made his controversial comments about rape. For me, that last straw threw the race into total confusion, and this may be the hardest Senate race to forecast because of it. Indiana remains a very Republican state on the presidential and gubernatorial level, yet the only public poll conducted after the rape comments showed a huge lead for Democrat Joe Donnelly. While I'm not sure I believe in such a huge margin, the momentum of the race is clearly away from the Republican—so I think this one will go blue too.

The one tossup where I ultimately expect the Republican to triumph is Montana. Democrat Jon Tester actually leads in the polling average at the moment, but this is one case in which I could easily foresee an upset, à la Nevada or Colorado in 2010. Montana is obviously a Republican-leaning state, and its support for Romney will present a hurdle for Tester; it's always harder to convince people to split their tickets than to vote the straight party line. This is, of course, the same reason why I chose Rick Hill to win the gubernatorial race in Big Sky Country. Indeed, my predictions are banking on Montana rediscovering its conservative roots.

Speaking of upsets—remember that, unlike the over-polled presidential race that I'll get to below, the smaller number of polls in many Senate races means there are fewer data available to make prognostications. This increases the margin of error and can make surprises more likely. That's a big part of the reason why there is usually an upset or two in the Senate races (and always a handful of upsets in the House). While I've picked them to go a certain way, this year I could see surprises taking place in Missouri, Arizona, and Nevada.

Missouri is an increasingly red state with an increasingly obtuse Republican candidate. Polls have shown Democratic Senator Claire McCaskill with a large lead because of it, but I've wondered for a while now if there may be latent support for Republican Todd Akin. No one wants to tell a pollster they're voting for an ostensibly sexist candidate, but when they get in the voting booth all by themselves, perhaps they'll fill out that secret ballot differently. This possibility is enhanced by the fact that most undecided Senate voters in Missouri are die-hard Republicans. In fact, according to Public Policy Polling, if Akin can pick up just 10% more Republican support, he wins. The libertarian candidate is also winning 6% of voters in the latest PPP poll, most of them Republicans. Will they come home? I do think it'll be close.

In Arizona and Nevada, the case is less complicated, and it all comes down to turnout among the Hispanic population. Democrat Richard Carmona is seeking to become the first Latino senator from Arizona, a state with a 30% Latino population but where Latinos made up only 13% of the vote in 2010. And, in Nevada, with an Obama win (spoiler alert!) looking likely, could the president's winning coalition pull Shelley Berkley over the finish line? Democrats in both states will be working hard to turn out these difference-making Hispanic voters.

In the end, though, I've still chosen to keep those three states in their "leans" categories. Totaling everything up, I project a 2013–14 Senate that consists of 53 Democrats (including two independents) and 47 Republicans—the exact same arrangement as today. Here's what it will look like:


Presidential


That takes us to the big kahuna. My view of the state of the presidential race should actually be pretty uncontroversial, at least to anyone who follows the polls and accepts their consensus. It feels as though we've settled on a nonet of swing states whose combined 110 electoral votes will decide the next president; furthermore, with the rich supply of polling we've had there, it is fairly easy to put them on a spectrum from blue to red (numbers are as of Sunday afternoon):


I do not consider any state apart from these nine to be worth investigating in the presidential contest. In the week before Election Day, we've seen political observers go even more cuckoo than usual over states like Pennsylvania, Minnesota, and Arizona. Despite last-minute candidate visits and polls showing a close race, I find it incredibly unlikely that any of these states will deviate from conventional wisdom. Historically, states that have been in one candidate's column all cycle long do not just suddenly defect to the other on election night. If Romney wins Pennsylvania, Michigan, or Minnesota, or if Obama wins Arizona, it will be because the national race on the whole shifted significantly toward him at the last minute; consequently, this entire prediction will be rubbish, since that will have meant one of the candidates won in a landslide. In the end, no state other than the nine in the chart above will be decisive in the Electoral College, because no other states will be swing states in an election consistent with how closely contested this one has been.

To walk through those nine states—it would appear that the ones on the extremes would be the easiest to predict, and I agree with this school of thought. I expect North Carolina to go to Romney without much difficulty; beyond the public polling average, the president has not visited the state since the Democratic convention two months ago, suggesting that his campaign does not see it as truly competitive.

In Nevada, early voting shows a huge advantage for Democrats, leading the state's most prominent political analyst to expect an Obama win. Mark Mellman, Harry Reid's pollster who was one of the few who accurately forecast that Reid would hold on in 2010, also has Obama up. Then there are just the demographics: as long as the state's growing Hispanic population prefers Democrats, Nevada as a blue-leaning state may be the new normal.

Finally, Wisconsin is a state where the Romney/Ryan ticket has invested a lot of time and energy, keeping it within their reach. However, past election results show that Wisconsin just isn't anything but a blue state, and even a native vice-presidential candidate can't change that. There hasn't been a poll showing Romney leading in Wisconsin since August 19 (before the Denver debate and his greatest moment of the campaign), and the odds of a candidate winning a state despite so many data points to the contrary are astronomical. This same logic can be applied to North Carolina and Nevada; hence, I feel pretty confident about my predictions in this first triad of swing states.

Skipping Ohio for a moment, the next-closest states are our old friends Iowa and New Hampshire. These states have both sported the occasional polls showing Romney with a lead, though Obama still holds the upper hand on average. They also lack the solid additional circumstantial evidence of an edge one way or another that Nevada, Wisconsin, and North Carolina can boast. Given their historical proclivity for being very elastic, fickle states, I could certainly see them going for Romney, but I still have to side with the averages and say they'll fall in the blue column, where the polls have them pretty comfortably resting for the time being.

Totaling my predictions so far from above, we stand at 263 electoral votes for Obama and 206 for Romney with four states left to examine—any one of which would put Obama over the top. It's the same argument you've been hearing for a while now, so I apologize for repeating it, but it's true: Obama is the favorite in this election because he simply has more paths to 270 electoral votes. This election is like a best-of-seven playoff series where the incumbent leads three games to none. Historically, it's a commanding lead, with the blue team needing to win only one of their next four games. But as Mitt Romney's favorite team, the Boston Red Sox (who did it while he was their governor), have shown, it's possible to win four in a row.

But I'm trying to give you my for-entertainment-purposes-only pick, not a synopsis of the odds. Indeed, this is where this presidential forecast verges from grounded in fact and reason to based on the flip of a coin. In contrast to the states above, which I expect to be very close but resolved on election night, the next three are the ones where I don't think we'll know the winner until sometime during the day on November 7. (Keep a particular eye on Colorado, where Secretary of State Scott Gessler's controversial tactics and malfunctioning voting machines have made the state ripe for a legal battle.) In other words, these are the pure tossup states, the ones that will be decided by the decimal place.

My (arbitrary) pick in Colorado and Virginia is President Obama. Call it a hunch—and you'd be right—but I see a number of intangibles in his favor. Most importantly, there is Democrats' superior get-out-the-vote effort; Obama has more than twice as many field offices in Virginia as Romney does, and in Colorado "one top GOP consultant who has worked on presidential campaigns told [the Atlantic] he mentally added 2 to 4 points to Obama's polls in the state based on superior organization." Strong field teams not only are better at turning out a candidate's likely voters, but they're also more wont to tap into the "unlikely" voter pool—which polls of "likely voters" often miss. Registered voters who can't reliably be expected to make it to the polls usually skew Democratic and can include younger as well as non-white voters. Specifically, as with the Senate races in Nevada and Arizona, I believe Hispanic voter turnout could be a difference-maker in Colorado. You can bet that OFA will be trying to get every last vote out of the state's Latino community, which polls can often undersample and therefore underrate the impact of. Finally, one last group that many polls miss entirely are "cell-phone-only" voters—that is, voters, likely or not, who do not own a landline phone. Because many polling companies do not call cell phones, their results may underestimate this blue-leaning demographic.

These are all acceptable justifications for picking Obama when all other factors seem equal. But there are also similar arguments that favor Romney—perhaps the most convincing of which is to point at the national polling numbers. Unlike the state polls, where Obama has had an advantage, national polls have been consistently better for Romney, showing either a tied race or the Republican leading. (For the record, I do not foresee a popular vote/Electoral College split. That's not something a responsible predictor can ever "call," in my mind, because it has been so rare historically and the election would have to be uniquely close.) You can average the national polls with the aggregate of state polls to tell a more conservative-friendly tale than the state polls alone show. This is as good a reason as any to break my mental tie in the state of Florida in favor of Mitt Romney. Another is the advantage he seems to have in the swingiest region of this swingiest state: the I-4 corridor between Orlando and Tampa. With its high percentage of conservative Cuban-Americans, Florida is also a state where high Hispanic turnout would not necessarily aid Obama.

Finally, we return to Ohio—and, unfortunately for the GOP, this is the real stake in Romney's heart. Even if Romney wins all three coin flips above, he'll still need Ohio for the presidency, like every Republican before him—but as we saw from the chart above, Ohio is somewhere between Iowa and Nevada in terms of how safe it is for Obama. This is a state where liberals have been reenergized thanks to a concerted union effort to kill a ban on collective bargaining, where liberal senator Sherrod Brown is poised to cruise to reelection, and where you can bet Obama's ground-game advantage will manifest itself as much as anywhere (if you're an OFA volunteer from a non–swing state, you're going to Ohio.) And then there's the stubborn polling advantage that has implied that Obama's support is as solid as a rock. As Nate Silver wrote of Ohio, "There are no precedents in the database for a candidate losing with a two- or three-point lead in a state when the polling volume was that rich."

Because of Ohio, and also simply because of Romney's dire reliance on my three coin-flip states versus their expendability to Obama, I am confident that Obama will win a second term. I am much less confident about my specific prediction that he will win 303 electoral votes to Romney's 235, but there you have my map anyway:


Now let's see what happens.

Monday, October 29, 2012

In Defense of Nate Silver

It's Nate Silver's job to analyze the news—so it must have come as quite a shock to him today to find himself become the news. While criticism of Silver has been out there for a long time, its most recent form has cut straight at the heart of Silver's analysis and represents the same type of anti-intellectual fear that has followed trailblazers like him around for centuries. In a surprisingly acidic POLITICO article, Dylan Byers makes Joe Scarborough's case against Silver and his data-driven polling analyses:
"So should Mitt Romney win on Nov. 6, it's difficult to see how people can continue to put faith in the predictions of someone who has never given that candidate anything higher than a 41 percent chance of winning (way back on June 2) and — one week from the election — gives him a one-in-four chance, even as the polls have him almost neck-and-neck with the incumbent."
Critiques of this ilk betray an inability to even speak intelligently on the subject of statistics, let alone a leg to stand on when presenting a counterargument to the findings of Silver's trademark Electoral College–predicting model. As I write this, Silver and his model give Barack Obama a 74.6% chance of victory on November 6. That number is very prominently labeled on Silver's website as "chance of winning." There's not much ambiguity in that. It should be obvious to anyone looking at that figure that what that means it that, in the judgment of the model, Mitt Romney has a 25.4% chance of winning the presidency.

But Byers, in the passage quoted above, clearly misses that point. Nowhere does it say that those 74.6%-to-25.4% figures are a prediction that Obama will win or that Romney will lose. It is an attempt to take a snapshot of the data and figure out odds. As Silver told Byers in the POLITICO article, there is still a significant chance that Romney wins—indeed, specifically, a one-in-four chance. If Romney wins, the model was not necessarily wrong. Indeed, every fourth time the model was run (Silver runs 10,001 simulations per day), Romney did win—and it's not a contradiction to say so while still handicapping Obama as the favorite.

Scarborough and, apparently, Byers seem to have a problem with this, but they don't seem to understand that this is the scientifically responsible way of doing this sort of thing. There is an academic discipline known as statistics, and they've been doing this a whole lot longer than any of us. Silver and others trained in this fickle art adhere to time-tested tactics such as the scientific method, gathering as-large-as-possible sample sizes, and acknowledging and even embracing the possibility of error.

In a world of post-debate insta-polls and Senate race rankings that are either Lean Democrat or Lean Republican, we as a society place a huge emphasis on "calling" states, elections, World Series, you name it. Audiences want instant gratification, and pundits give it to them with iron-clad predictions that they finalize and stick to come hell or high water.

What makes Nate Silver so unique—and so valuable—is that he resists that entirely (and yet still manages to be popular; imagine that!), favoring instead a scientifically responsible spectrum. The core tenet of this method lies in the difference between a 49% chance of an Obama win and a 51% chance of an Obama win. For most pundits, those are opposite predictions. On a spectrum, they're virtually identical. Given that it only takes a two-percentage-point swing to make up that difference, that's the right way to think about it.

Likewise, a spectrum always leaves room for some doubt. Even very safe predictions have a small chance of not happening, and a probability spectrum is honest about that fact, setting 99% or 99.9% odds for a very likely event. In other words, a good scientist always leaves room for the possibility that anything from extreme X to extreme Y will occur; the trick of creating a utile spectrum is knowing where to fix the "tipping point" between "lean X" and "lean Y," not picking one or the other. The beauty of a good probability spectrum is that it allows for every possibility. That's because the chance always exists, however small, that something extremely unlikely (e.g., a Romney landslide) will happen. In that sense, spectra like Silver's model will always be accurate.

And maybe that's the problem; skeptics see Silver's model's tolerant spectrum as wishy-washy—an attempt to take credit for being accurate no matter what the outcome. Science has one word for these people: "Tough." We have no choice but to accept this little ambiguity in our lives, because we have no way of ever being certain about anything. I understand that that is unsettling for many people, but that's what being a scientist—or even just being intellectually curious—is all about.

It also doesn't help matters that the exact figures and contours of a probability spectrum are impossible to prove. No one can ever say for sure that, on October 29, Barack Obama had a 74.6% chance of winning the race, even if Romney does win in that landslide. All we will know is the binary outcome: did Obama win or not, and by how much. It takes a much broader body of work to "prove" (to the extent anything can be proved) that those odds were correct—a body of work that, sadly, we'll never have. (You'd need the 2012 election to duplicate itself in future elections exactly the same way through October 29 a few hundred times, then see who won in each of those cases. In laboratories, these types of experiments are possible. Not in political science, where this is only the 57th presidential election in American history.) The best Silver—or anyone mortal in the whole wide world—can do is make an educated guess based on the data we do have. You may criticize which data—which polls or which economic variables—get plugged into Silver's model; you may not ignore science or the discipline of statistics.

Yet people do. People rely on their "gut" more than on the data in many more fields than just politics, and Silver has been dealing with them his whole life. As an early employee of Baseball Prospectus, Silver invented the PECOTA system and was an early figure in baseball's sabermetrics. He and co-Moneyball-ers tried to bring a rational, data-driven approach to predicting baseball the same way he has done in politics—and met with the same uninformed ridicule.

Baseball is full of the same "anti-statheads" that have come out of the woodwork in politics recently. You know them as the people who think of pitchers' wins as still a valuable statistic. They're the ones that denigrate WAR by saying that a better measurement of skill is actually how many wins you generate above a replacement level. They believe in momentum in baseball, in "clutch" hitting, and in the idea of lineup protection just because their experiences have led them to.

(Note: I'm painting with an extremely broad brush. In fact, I would like to see more rigorous statistical study on each of those last three. And you can indeed have a reasonable argument with other baseball experts or fans about those things—as long as the argument is empirical and grounded in data and facts, not "general impressions.")

The anti-Silverites in politics we see today are the descendants of the meanest versions of that baseball old guard: the old-timey scout who believes stats and innovation have nothing to offer him; the longtime columnist who bullies and mocks statisticians as "eggheads" or "binder boys." These people are as closed-minded as Nate's probability spectrum strives to be open-minded. The best analysis, and the best predictions, will inevitably come from viewing all available data and considering them holistically. As HardballTalk head blogger Craig Calcaterra says, quite astutely I think, if you worked in any field other than baseball and stubbornly ignored new information and new technology in your job, you'd be fired. Any field other than baseball or politics, I guess.

Maybe it's my wishful thinking, but it seems to me that those people in baseball are, fortunately, becoming more and more marginalized. Unfortunately, though, that's what makes those critics in politics much more dangerous—they are actually "important" pundits who are taken seriously. Indeed, in baseball, people can be as ignorant as they like, but the only real damage they're doing is taking up column inches and maybe, just maybe, encouraging a stupid trade to go down. In the powerful field of politics, ignorance can have a real effect on policies or the next leaders of the United States. They're playing with fire.

That's all the more reason to make sure Silver's voice of reason isn't drowned out. Unfortunately, Nate has caught onto the fact that many people in politics are jerks, and it may hasten his "retirement" from the field of political forecasting. (This most recent incident can't have helped.) But with Nate gone, unlike in baseball, the statistics-ignorant crowd will have won out, and political observers and the viewing public will go on thinking that the tools of ignorance are an acceptable forecasting model for elections.

I happen to roughly agree with Nate's prediction on the outcome of the presidential race, but losing his crystal ball is not even close to the reason his departure would sting. Rather, it's the loss of the reasonable, data-driven approach that he represents and has brought to the fields of baseball, politics, and others. Silver stresses that prediction is an imperfect science that he's just trying to make sense of, not solve. He is a student of the science of prediction, not a prognosticator per se. As he likes to say, it's not about which predictions are right, but rather which are less wrong. Instead of trying to eliminate error like those pundits and their iron-clad predictions, he truly does embrace it and see its value in helping improve subsequent forecasts. He's enhancing the study of predictions and helping us understand how to make them better—a goal that's bigger than elections, in some cases even helping to save lives.

Anyone who has given Silver's New York Times blog a more than cursory read recognizes all this, because Nate goes to pains to point it out (undoubtedly stung by ignorant would-be statisticians before). This isn't weakness, or being "unmanly," as it was distastefully put this week. It's realism and nuance, two traits that are essential for level-headed people in any field—only when it comes to predictors, they're basic qualifications.

Tuesday, October 23, 2012

Introducing the Baseballot Gubernatorial Rankings

With only two weeks left to go before Election Day, I considered the other day the relative lack of attention that governors' races are getting this cycle. This is hardly a new phenomenon for the relatively few gubernatorial elections that take place in sync with the presidency, but it is an unfortunate one. Governors, along with state legislatures (whose electoral prospects are fascinating but, alas, I will not have time to address before November 6), affect people's day-to-day lives in the states in question arguably more than the Senate and president will. Accordingly, I wanted to devote some more space on my little corner of the internet to assessing the 11 gubernatorial races of 2012.

At the top of the page, you'll see a new tab, "Gubernatorial Rankings," added to the menu. Click through to see my race rankings for the 11 campaigns in the same chart form as my Senate rankings. The scale is also the same—Solid/Likely/Leans for each party, explained in detail here.

Because of the relatively few number of gubernatorial races, however, I figured I could devote a little bit of time to a qualitative analysis of each one as well. In alphabetical order:

Delaware (Solid Democrat)
Sometimes, less is more. Democratic Governor Jack Markell is a popular governor in a blue state in a presidential year. His Republican opponent is not a big name. Easy hold.

Indiana (Likely Republican)
This could just as easily be Solid Republican, since there's no polling evidence that Democrat John Gregg is competitive against Republican Congressman Mike Pence. However, Gregg is a solid candidate who has been able to get active on the airwaves and has achieved some fame for his crazy moustache. With the competitive Senate race in Indiana, this at least could be closer than the GOP blowout it otherwise would be.

Missouri (Likely Democrat)
Governor Jay Nixon has proven to be a very strong candidate for the Democrats, winning over solid crossover support despite a rocky beginning to the cycle. Furthermore, whether hurt by association with Todd Akin or just falling flat of his own accord, Republican Dave Spence—oddly, in my opinion—hasn't been able to tap into Missouri's growing GOP base of support.

Montana (Tossup)
Sparse polling hasn't given us many clues about who's favored in this race. Montana is a reliably red state in the presidential race, but a close race for Senate and its love for outgoing Democratic Governor Brian Schweitzer has made it swingier this cycle. Schweitzer afterglow would probably be responsible if Republican Rick Hill loses this one to Democratic Attorney General Steve Bullock.

New Hampshire (Tossup)
A true swing state on every level (president, Senate, both House seats, and governor in 2012), New Hampshire has a popular Democratic governor who is retiring. I would have thought that that fact, plus a weak (in my opinion) Republican candidate in Ovide Lamontagne (he was Kelly Ayotte's Tea Party nemesis in 2010), would create an advantage for Democrat Maggie Hassan, but polling so far has been tight—with the two candidates seemingly playing hot potato with the lead with every poll that's released.

North Carolina (Likely Republican)
North Carolina may end up the only state where the corner office changes party control in 2010. Republican Pat McCrory is well positioned to win thanks to an unpopular Democratic incumbent, up by double digits in most polls. Ironically, though, Democrat Walter Dalton's status as lieutenant governor—meaning he's won statewide before—is the only thing preventing me from ranking this as Solid Republican. Wild card: will Romney's withdrawal from North Carolina hurt McCrory?

North Dakota (Solid Republican)
Call it Delaware in reverse: Republican Governor Jack Dalrymple is very popular, and an unknown Democrat is just not going to overcome him and Mitt Romney.

Utah (Solid Republican)
One of the reddest states in the country electing its governors in presidential years? No wonder Utah hasn't elected a Democratic governor since 1980. Republican Governor Gary Herbert is safe.

Vermont (Solid Democrat)
Vermont, along with New Hampshire, elects its governors every two years. While Democratic Governor Peter Shumlin won office in a squeaker in 2010, this year will be much friendlier to Democrats. That's been borne out by polling showing Shumlin clearly in the lead over Republican Randy Brock.

Washington (Leans Democrat)
This is the heavyweight battle of 2012—Democratic Congressman Jay Inslee in a blue state against moderate Republican and sitting Attorney General Rob McKenna. Washington hasn't elected a Republican governor since 1980, but if anyone is going to do it, it'll be the extremely strong and likeable McKenna. He led in pollsfor much of the spring and summer, but as voters have been reminded of the presidential race and their preference for national Democrats, Inslee has taken a slight lead. While most of his leads remain in the margin of error, various pollsters have all been consistent about saying Inslee is the one with the edge lately, so I will too.

West Virginia (Likely Democrat)
This seems like it should be safe for Democrats, as current Governor Earl Ray Tomblin was able to win a special election in 2011 when the national mood was still very anti-Democrat. However, there has been no polling in the race, leaving it a bit of a dark horse. West Virginia has identified with the Democratic Party for decades, remaining faithful to it on the state level even while voting consistently for Republican presidential candidates. Tomblin is the right kind of pro-coal Democrat whom Romney voters won't hesitate to support, but Republican Bill Maloney should have to only subtly shift Tomblin's public image to make him look much more like one of those hated "Washington liberals." If that happens, you could see a 20% shift of voters from Tomblin to Maloney en masse. Right now I see Tomblin winning this race by about 10 points, but the tables could suddenly turn. One sure prediction: in this presidential year, it won't be a close one; the most Democratic Romney-compatible candidate should be able to dominate.

Monday, September 10, 2012

Introducing the Baseballot Senate Rankings

Too much attention is paid to the presidential race, and the dozens of polls in each swing state make it too difficult to create any stable kind of ranking. The race for control of the House, meanwhile, has far too many unknowns and, with 435 races, is a far too ambitious project for only eight weeks out. Therefore, here in 2012, in Baseballot's first election cycle of existence, I have decided to issue my personal rankings for the race for control of the US Senate, which are now available by clicking the appropriate tab above.

I initially issued these rankings over Twitter a couple weeks ago; there has been only one change, which I also tweeted about today. That first announcement (a soft launch, if you like) prompted a good round of debate and discussion with some followers, and I encourage readers to do the same thing with the rankings now that they are posted in full on the site.

Bringing the rankings off Twitter and into a somewhat more institutionalized form allows me to do a few things. First, it lets me display them in a more creative way—at least, as creative as my Excel-chart-making skills will allow. Second, it lets me explain my thought process to you in more than 140 characters.

As you'll see by clicking the link, each race is organized into a row and a column. The columns represent which party currently holds the seat; the rows represent my ranking of the competitiveness of that seat. As a result, you can see at a glance (a) which party stands to win the seat in question and (b) which party may potentially stand to lose it. In this case, the chart demonstrates how the "tipping point" of competitiveness lies squarely on the Republican side; that is, there are no Republican-held seats where Democrats have the advantage, while Republicans are looking good in some of the races in the left-hand column. In other words, Democrats are playing defense.

One of the most important features of the chart is the fact that, while the x-axis (current ownership of each seat) is necessarily a binary, the y-axis is more of a continuum. Specifically, I have ordered the races, from top to bottom, from most Democratic-leaning to most Republican-leaning. This is because, despite the usefulness of labels such as "likely Republican," they can also be misleading—these things are on more of a "percentage-likelihood-of-winning" sliding scale. Put another way, I do not believe that Wendy Long (R-NY) has the same odds of winning as Pete Hoekstra (R-MI), despite the fact that they are both in the Solid Democrat category.

This has another useful visual side effect. If and when I update the chart, you might see some of the races move up or down in the ranking—or you might instead see the rankings move up or down the chart. The former would be if an event shook the odds in a specific state—for example, if the Republican candidate makes an offensive comment about rape. The latter would occur if the national electoral landscape appears to move in a more Democrat- or Republican-friendly direction.

As for my methodology—I regrettably do not have the numerological chops of Nate Silver, who in my mind has the last word in electoral rankings at any level and whose predictions and analysis I respect greatly. As a result, my rankings go more with my gut than with any statistical formula. (Naturally, however, my gut is informed significantly by polls, my assessment of the quality of those polls, and other both qualitative and quantitative race analyses that I read in publications and online.) You'll notice that I've used the widely accepted format of "Solid/Likely/Leans." To me, "Solid" means there is simply no evidence that that party can lose that seat. "Likely" means that I acknowledge that the seat is in play, but in my heart of hearts I don't see it going to the other party. "Leans" means I believe a race could reasonably go either way, but I have identified enough of a tilt that I'm comfortable making a call. And, finally, "Tossup" means I simply have no freaking idea.

I will make one observation about my formulation of the rankings. You may come to notice, as I have, that I tend to be conservative in my assessment of each race. By this I do not mean that, in a fit of either optimism or pessimism, I tend to be more bullish about Republicans' chances; I mean that my rankings tend to take the long view and be slow to change. I will not be adjusting the chart if one poll comes out showing a drastically changed race; I will treat it as an outlier until more polls prove otherwise. Even if a few polls in a given week show the complexion of a campaign to have changed, I'll usually hold off in switching the ranking until I can confirm that it wasn't just a temporary "bounce." I tend to believe that the picture painted over a broader sample size, on a more consistent basis, is the more reliable one. Indeed, I'll even take this to the extreme sometimes—if polling is showing a neck-in-neck race in a state that usually displays a strong partisan tilt, I'll be skeptical that those voters are really leaving that partisan past behind. This is why, for example, despite most polls showing Scott Brown in the lead, I continue to consider Massachusetts a Tossup; in a state where President Obama is likely to hit 60% of the vote, I simply consider it a tall order for Brown to win over 10% worth of ticket-splitters. (The reverse of the Massachusetts case is true in Indiana and North Dakota, which I see as Leans Republican but which polls show as tied.)

For those of you not curious about the individual subtleties of each race, I've provided the rankings' bottom line in the final column on the right of the chart: the number of projected Democrats, the number of projected Republicans, and the number of tossups. Currently, I foresee—as I have for a few months now—a complete tossup for control, with 48 seats looking blue and 48 seats looking red. That leaves four Tossups, which, if you take the definition of "tossup" literally, figure to split evenly between the two parties. That creates a Senate tied up at 50, with the future vice president (currently likely to be Joe Biden) the tie-breaking vote. (Given the conventional wisdom that Angus King will caucus with the Democrats, I've included him in the Democratic total for now, but it is certainly possible that in a truly tied Senate he would sell his services off to the highest bidder.)

Anyway, as has been said, the rankings have been added as a permanent addition to the site via the tab above. Check them often, as I anticipate they will change, and enjoy!

Tuesday, August 21, 2012

Reading Way Too Much Into the Latest Missouri Senate Poll

Despite their feelings about "The Wave," Public Policy Polling has a well-deserved reputation as the most "fan-friendly" pollster around. By this I mean that they have an active and amusing Twitter presence; they're generous and transparent with their data; and they give the people what they want, most recently turning around a Missouri Senate poll only 24 hours after Republican candidate Todd Akin claimed that "legitimate rape" doesn't result in pregnancy. PPP's poll, conducted Monday night, found that Akin remained ahead of incumbent Democratic Senator Claire McCaskill 44% to 43%.

Prior to Akin's mega-gaffe, my personal assessment of the race had been that it leaned Republican. After his comments, I commented on Twitter that I believed the race had shifted—to "teeters Republican," reflecting my opinion that the fundamentals in Missouri still favor Republicans but that Akin had brought them precariously close to the brink of defeat. I've since been surprised by the volume, speed, and intensity of the wrath that has come down on Akin, as well as the boldness of predictions from highly credible pundits that the incident would dramatically shift the electoral landscape in McCaskill's favor. In finding a virtually tied race, then, the new PPP poll—while hardly the final word on the matter—was reassuring in the sanity it seemed to restore. Gaffes—short of performing the Nazi salute—don't just erase candidacies overnight.

In their blog post on the most recent poll, PPP rightly points out that Missouri's, and the country's, increased polarization means that even major campaign revelations have a minimal impact on the horse race; voters are simply too dug into their respective sides to budge. (We see that with the remarkably consistent numbers in the presidential race too.) A direct comparison to the last time PPP surveyed a McCaskill/Akin head-to-head reveals almost no change—from Akin 45%, McCaskill 44% to the current Akin 44%, McCaskill 43%.

However, that poll was conducted in late May—before Akin's gaffe, sure, but also before the Republican primary, before the health-care ruling, before a lot of things. So that does leave plenty of room for the possibility that there has been volatility in the meantime—specifically, that Akin built himself a solid lead after May, only to see it evaporate in the past couple days after shooting himself in the foot. While we can never know for sure, a dive into the crosstabs of the two PPP polls provides some clues.

First, what may be most obvious (and what PPP themselves point immediately to) is Akin's favorability rating. Democrats and liberals dislike him and always have—but among McCain voters, his favorability went from 34/10 in May to 40/39 today. In other words, even Republicans who are getting to know this guy aren't liking him. However, they are still voting for him, at a 70/10 clip. This is hyperpartisanship in play and suggests that, while the incident didn't come without an effect, it isn't necessarily translating into people's votes.

To see movement in that arena, a better crosstab to look at is gender; Akin's comments were, after all, consistent with the "war on women" meme. Back in May, Akin didn't suffer from much of a gender gap. Among women, he trailed McCaskill 45% to 43%; among men, he was ahead 46% to 44%. Yet today, unsurprisingly, tells a very different story. If women were the only voters, Akin would currently lose to McCaskill 49% to 39%. However, Akin is up 50% to 36% (!) with men, even after Legitimate-Rape-Gate.

While I have no trouble believing that the cratering support of women is linked to this scandal, it is doubtful that Akin's "legitimate rape" claim actually boosted his support among men. (Interestingly, men were actually more likely to say that those comments were inappropriate and to "strongly disagree" with them, according to the survey—by about 10 percentage points!) Therefore, something else must have changed in the months between polls.

The simplest theory is that, sometime between May and August, Akin's support among men ballooned while his support among women did not—then, this weekend, his support among women took a major hit, resulting in the disparity we see today. Indeed, given the gender splits on the appropriateness of Akin's comments, it's very possible that his already excellent statistics with men were even better before this controversy. If you assume that his support among women had remained split evenly down the middle until the "legitimate rape" incident—a reasonable assumption, given that those stats moved in the "right" direction between polls—that would be enough to have given Akin a sizable overall lead had PPP polled the race just a week or two ago. While they are different polling firms and a direct comparison is impossible, this jibes with the picture of the race painted by Mason-Dixon in late July. That poll found Akin ahead of McCaskill 49% to 44%.

Maybe, then, Akin did take a statistically significant hit from his controversial comments about rape—and if you're a Republican, it has to be worrisome that that's just after one day of fallout. But since we made an assumption that the gaffe accelerated Akin's current gender gap, this may just be an instance of a mathematical identity. What actually seems like a safer conclusion to draw is that, whether Akin has experienced a drop in the past few days or not, he does not have much further to fall. The political crosstabs reveal that Republicans are loath to abandon their candidate, even if they despise him, and that McCaskill can't count on picking up more support there. Meanwhile, the gender crosstabs reveal that virtually every non-Republican woman is already voting against him. Other, traditionally strongly Republican slices of the population will have to start abandoning Akin for his support to fall below its current threshold. Unless McCaskill can start making inroads with men and McCain voters, Akin should continue to be viewed as extremely competitive.

Tuesday, August 7, 2012

Bleeding Kansas

Electoral success can be a double-edged sword. If a political party is too dominant—to the point of pushing out the other major party (and therefore the other major dissenting voice)—new disagreements and new fissures inevitably pop up. In the worst-case scenario, it can lead to all-out intra-party war.

This is the situation that the Kansas Republican Party finds itself in today. In that state, Republicans control both US Senate seats, all four US House seats, the governorship, 92 of 125 State House seats, and 32 of 40 State Senate seats. However, this seemingly advantageous position has given way to an irreparable schism in the party between moderate Republicans, who claim the "establishment" mantle, and conservative Republicans, led by Governor Sam Brownback.

It has gotten so bad that, despite Republicans' overwhelming numbers in the State Senate, it is actually the progressives and moderates who control the chamber—a European-style coalition of the body's eight Democrats and a handful of moderate Republicans who have evidently concluded that they share more in common with the average left than the hard right. Indeed, in Kansas, party labels don't even begin to tell the full story—something unique in American politics today.

With the moderate coalition in the State Senate the only obstacle to enacting Governor Brownback's deeply conservative agenda (e.g., union busting and spending cuts), the intra-party schism has turned ugly—a modern political equivalent of the pre–Civil War Bleeding Kansas. Enraged conservatives, accusing their moderate colleagues of disloyalty to the cause, have actively tried to purge their fellow Republicans from elected office. Similar to the establishment-versus–Tea Party dynamic in other states, it has resulted in many moderate-versus-conservative Republican primaries for State Senate across the Sunflower State. Yet Kansas has taken it to a degree that the Tea Party wouldn't dare; the bizarre scenes include labor unions draping their arms around moderate Republicans, Democrats actively campaigning and voting for the same, and conservatives attempting to tie "moderates" (who are still probably to the right of John McCain) to "socialist" President Barack Obama. Things are so heated that state senators representing roughly 70,000 people each are raising millions of dollars. (It doesn't hurt that the Koch brothers are based out of Kansas.)

It all comes to a head tonight—primary election night in Kansas. Because Republicans are overwhelmingly favored to win in most districts, and because Kansas state senators serve non-stacked four-year terms, tonight will determine whether the moderates or the conservatives control the Kansas State Senate until 2017. For Kansas residents, the stakes are high; for the rest of us, the state provides a fascinating view on the consequences of one-party control and the extremities to which the current trend of polarization can take us.

Polls close in Kansas tonight at 7pm local time—that's 8pm Eastern in most of the state, but 9pm Eastern in a small sliver on its western edge. Only the State Senate, where there are currently eight Democrats, 16 moderate Republicans, and 16 conservative Republicans, is in danger of switching control (but not, of course, parties!).

Here are the races to watch to see which faction will seize control. Because these things aren't really black and white, it's estimated that three or four conservative gains would be enough for them to seize power.

Seats Currently Held By Moderates

District 7: Kay Wolf (moderate) vs. David Harvey (conservative). This is an open seat; Senator Terrie Huntington is retiring. Wolf, who is a state representative already, is probably the slight favorite.

District 8: Tim Owens (moderate) vs. Jim Denning (conservative). Owens is running for reelection and is a big voice for the moderates, so the conservatives will love it if they can pick him off.

District 11: Pat Colloton (moderate) vs. Jeff Melcher (conservative). This is an open seat being vacated by moderate Senator John Vratil. It will be one of the more competitive races in the state.

District 20: Vicki Schmidt (moderate) vs. Joe Patton (conservative). Schmidt is the current Assistant Majority Leader, but because Patton has support from prominent conservatives like Secretary of State Kris Kobach, this will be another nail-biter.

District 22: Roger Reitz (moderate) vs. Bob Reader (conservative). Senator Reitz is lucky that Reader is splitting the conservative vote with a third candidate, Joe Knopp.

District 24: Pete Brungardt (moderate) vs. Tom Arpke (conservative). Brungardt is the incumbent here, while Arpke is a current state rep.

District 25: Jean Schodorf (moderate) vs. Michael O'Donnell (conservative). The catch here is that, if incumbent Senator Schodorf loses, Democrats would have a shot at a pickup here in November.

District 26: Dick Kelsey (moderate) vs. Dan Kerschen (conservative). Senator Kelsey is in less danger than some of his colleagues, but this race has still been targeted.

District 31: Carolyn McGinn (moderate) vs. Gary Mason (conservative). McGinn has been in the State Senate since 2005.

District 37: Pat Apple (moderate) vs. Charlotte O'Hara (conservative). Incumbent Senator Apple isn't the most moderate Republican in the Senate, but he is moderate compared to the extremely far-right O'Hara. This district leans Apple.

District 39: Stephen Morris (moderate) vs. Larry Powell (conservative). Morris is the current Senate president; needless to say, his loss would be a huge blow to moderates.

Seats Currently Held By Conservatives

District 1: Marje Cochren (moderate) vs. Dennis Pyle (conservative). People generally aren't taking Cochren's challenge to the incumbent Pyle very seriously, but this is a more moderate district—you could see Democratic crossover voting.

District 10: Tom Wertz (moderate) vs. Mary Pilcher-Cook (conservative). Pilcher-Cook is an incumbent seeking her second term. She is favored to win.

District 40: John Miller (moderate) vs. Ralph Ostmeyer (conservative). Ostmeyer is the incumbent here, and while he's considered the favorite, Miller has given a lot of his own money to his campaign.

Seats Created By Redistricting

District 15: Dwayne Umbarger (moderate) vs. Jeff King (conservative). Umbarger and King, both incumbent senators, were thrown together in redistricting. Umbarger is generally expected to win.

District 21: Joe Beveridge (moderate) vs. Greg Smith (conservative). This is a new district, but the member whose retirement made its creation possible was a moderate. Conservatives have a good chance of picking up this seat.

Thursday, July 19, 2012

Politically, the Veepstakes Isn't Worth a Bucket of Warm Spit

I'm proud to count myself among the chorus of adherents to the scientific method even in non-scientific settings. To stereotype, in baseball, these people are the sabermetricians; in politics, they're the political scientists. Yet despite their rising profile and their many successes, their counsel and predictions are often ignored in favor of that of scouts and campaign operatives. Again, while this is a stereotype, these are the people who are more apt to draw false conclusions based on a perceived pattern from their years of experience.

We're treated to that every four years on the grandest scale, when at least one major party nominee chooses a vice-presidential running mate. Conventional wisdom among the media, political professionals, and basically everyone else holds that it's the most important strategic decision a campaign can make—capable of delivering the votes of entire demographics and, ideally, a vice-presidential candidate's home state in the Electoral College.

There is no hard evidence for this. What's more, there is no soft evidence for this. In fact, it has been demonstrated time and time again that there is no "home-state advantage" for a vice-presidential nominee's ticket. Consequently, while Mitt Romney may have his own very good reasons for picking Senator Marco Rubio or Senator Rob Portman as his number two, he should be careful not to do it thinking it gains him an edge in their respective home states (Florida and Ohio, two of the biggest swing-state prizes out there). Unfortunately, Romney is getting advice to the exact opposite effect. Not only does the media speculate endlessly about home-state advantage when it comes to a VP pick, but Republican insiders place a premium on it too. One Portman proponent drew the connection as starkly as it could be drawn: "He's from Ohio, and we need to win Ohio."

This assumption is fostered by a lot of anecdotal, and some scientific, evidence. Just this week, a CBS News/New York Times poll found that 26% of respondents claimed the vice-presidential nominee mattered "a lot" to their vote, and 48% said it would affect their vote "somewhat." Unfortunately for campaign operatives picking a VP, however, this is a long way from proof that their selection matters. It is one thing for people to say in the abstract that the bottom of the ticket is important; it is another to present them with different possible tickets (e.g., Romney-Jindal, Romney-Pawlenty) and see if that truly sways their opinion.

Indeed, few and far between are the polls that show an actual, concrete vice-presidential pick making a difference outside the margin of error. For instance, Fox News polled Ohio residents about a generic Obama-Romney matchup, but then asked about a more specific Obama-Biden vs. Romney-Portman matchup. The change between the two amounted to little more than noise in the data; the control resulted in a 45%–39% Obama edge, while the inclusion of the VP candidates made it 46%–40%. That's a one-point change for each candidate, and a net change of zero.

Similarly, in a poll of Florida, Public Policy Polling discovered that putting Rubio on the ticket actually decreased Romney's share of the vote by two points—from 45% to 43%. Worse, President Obama's share of the vote remained steady at 50%. Another rumored vice-presidential candidate, Jeb Bush, also barely registered an impact. He shaved one point off Obama's vote share and gave one point to Romney, for a 49%–46% race. And this was the most benefit given by any possible VP nominee that PPP tested! Not only does this show that a VP nominee changes very few minds, but it also pokes a giant hole in the belief that, even if the nominee does change minds, it provides an automatic bump in the nominee's home state. Yet the belief in the importance of the bottom of the ticket persists.

Granted, there are polls that show the opposite result—that is, a vice-presidential candidate making a difference. A May poll by Quinnipiac, for instance, was upheld as evidence that Rubio would help Romney eke out a win in Florida. But even that poll showed only a two-point boost with Rubio, with the article noting below the (e-)fold that that difference was within the poll's +/-2.4% margin of error. The point is that, for every poll showing the VP nominee making a difference, there is one that shows no effect—making the impact of vice-presidential selection basically inconclusive. Furthermore, if you can find me a poll that attributes a five-point bump to any vice-presidential candidate (and if you can, please send it to me), please also find for me the last time a poll in July accurately pegged the final margin of victory in November. Circumstances change, and literally hundreds of political events will intervene between now and then. It is a reach to assume that people reacting in July to the naming of a vice-presidential candidate carry that reaction with them for four months and have it continue to weigh on their vote on Election Day.

The reality is that people think that their vote is based partially on the VP half of the ticket. They think of themselves as making a thoughtful, balanced choice, considering many variables (of which the ticket is one)—so they tell a pollster that, yeah, it's somewhat important to them. (This is also the reason that political operatives think it matters.) But when they make that choice, how often is the vice-presidential nominee the tiebreaker? Or, realistically, how many people like most things about the Democrat but then decide to throw all that out the window because the Republican vice-presidential nominee is from their state? How much sense does it make to base your vote on the number-two guy over all of the factors to consider—which are about the number-one guy?

Indeed, most people who really like the number-two guy enough to be swayed to vote for him (or her) already like the number-one guy, too. In other words, they are already strong partisans or votes that are already in the bag—even if they didn't seem to be. I concede, for example, that a wavering evangelical could be reassured into voting for Romney by a solidly conservative pick like Chris Christie. (Indeed, these were the findings of a PPP national poll in April: "Christie's inclusion doesn't have a big impact with Democrats or independents but it helps Romney shore up the party base, going from an 82-10 lead with Republicans to an 86/9 one.") However, as Election Day approached, those waverers would have coalesced around Romney anyway, in a well-documented political-science phenomenon known as "coming home" to the base.

This overall theory lines up pretty well with the empirical findings of political scientist David W. Romero, who conducted one of the most definitive studies on the issue of the vice-presidential nominee and voting patterns. Reviewing the political-science literature, Romero found—as we did—that surveys of the individual voter gave the vice-presidential candidate a lot of weight, but that that impact disappeared when looking on the aggregate level—removing from the equation individuals' "rationalization" of their vote. He divided up the factors that influence electoral decision-making into two camps—those pushed from the campaign to the voter (e.g., ads) and those intrinsic to the voter (e.g., partisanship). However, he found that most of the campaign-generated persuasion was geared toward the presidential candidate. The little that was related to the VP candidate tended to be grouped around the summer, which another study tells us is almost irrelevant to November (campaign ads during the summer just don't have the "half-life" to continue to influence people months later). As a result, the only way the vice-presidential nominee can influence a voter is by appealing to his or her intrinsic factors; or, as Romero hypothesizes, "the voters’ evaluation of the vice presidential nominees is at best flimsy and impressionistic, and at worst simply a function of their predispositions." When he controlled for those predispositions in his study, he found that, no matter how much voters thought they were considering the VP nominee, it simply didn't switch any votes in practice.

If you don't trust Professor Romero, we need look no further than trusted election guru Nate Silver's simpler, more plain-English study concluding the same thing: that the vice-presidential nominee barely moves the needle in his or her home state. Silver took just about the most straightforward approach you can to this problem: he looked at how well tickets did in the home states of their VP nominees relative to their nationwide performance. His findings averaged out to a 2.2% bump for the hometown VP's ticket—not bad, but certainly not a game change. In addition, the historical range fluctuated everywhere between a 21.1% bump (John Nance Garner in 1932) and a 17.8% disadvantage (Joseph Taylor Robinson in 1928). In other words, the effect is wildly unpredictable.

Yet despite this mountain of evidence, there continues to be a massive disconnect between reality and perception when it comes to the selection of a vice-presidential nominee. That nominee almost never provides the requisite boost to put a state over the top for his or her ticket, and that state has almost never been the difference-maker in the Electoral College. Still, however, prognosticators today continue to talk about swing-state status as the hot-button issue in the veepstakes. Can they actually believe this to be true? With all the evidence coming down against it, and the almost complete lack of proof to support that belief in the first place, there is absolutely no reason for anyone to believe that a vice-presidential pick makes a difference in how people vote.

No, votes for the presidency tend to boil down to the very things you'd expect them to: partisanship, ideology, presidential approval rating, and, especially this year, the state of the economy. When it comes to Mitt Romney's choice of a running mate, the best advice for any presidential campaign continues to be, "Keep it simple, stupid!"