About six weeks ago, Slate launched Lean/Lock, an interactive game that allows readers to test their skills as a political pundit. With one month to go until the election, we decided to take a quick look at how your predictions align with those of some more prominent forecasters.
In Lean/Lock, players can predict the winners in 36 Senate, House, and governors' races and make their picks provisional ("leaning" toward a candidate) or certain ("locking" on him or her). So far at least, Slate readers don't radically disagree with the professional pundits. But they have developed enough differences to make a meaningful comparison. If you throw out what the forecasters call "tossup" races, aggregate Lean/Lock picks diverged about 15 percent of the time from the predictions of four traditional forecasters—CQ Politics, Cook Political Report, Rothenberg Political Report and Larry Sabato's Crystal Ball. Lean/Lock players disagreed more strongly, however, with Nate Silver's simulations at Five Thirty Eight: Overall, players disagreed with 66 percent of Silver's predictions. Here's a look at a few races where the game has produced curious results.
Slate readers say: Rand Paul (R) 57 percent, Jack Conway (D) 43 percent. Thirty-three percent of picks are locked.
Pundits say: Leans Republican (Sabato, CQ); tossup (Rothenberg, Cook).
The Lean/Lock database shows that Conway led Paul for most of August, initially by 20 points. But Paul, a libertarian by philosophy if not party, began gaining in late August and now leads by 13 points. The situation on the ground in Kentucky appears to tell the same story: Voters initially found Paul too extreme to be a serious candidate but have warmed to him considerably as he has inched closer to the political mainstream. Two-thirds of the players are keeping their options open, however. The blip around Sept. 18 initially confused us, since news searches in that window don't turn up anything bombastic. A look under the hood shows an influx of new players around that time that briefly threw off the standings. They normalized within a few days.
Slate readers say: Joe Miller (R) 86 percent, Scott McAdams (D) 14 percent. Fifty-three percent of picks are locked.
Pundits say:Likely Republican (Sabato, CQ); safe Republican (Rothenberg, Cook).
From Day One, the vast majority of players have considered Miller a shoo-in, though Lean/Lock doesn't allow for third-party candidates like incumbent Sen. Lisa Murkowski, who lost the Republican primary to Miller. Her presence in the race could explain why many players have yet to lock their picks. Miller's standing in Lean/Lock almost exactly matches the odds that 538's Nate Silver, using a statistical modeling approach, places on Miller.
Slate readers say: Raul Labrador (R) 80 percent, Walt Minnick (D) 20 percent. Forty-two percent of picks are locked.
Pundits say:Tossup (Sabato); leans Democratic (CQ, Rothenberg, Cook).
The extreme confidence in the Republican is a bit puzzling here. But House races are harder to predict than statewide races, and the fact that fewer than half of those predicting Labrador's victory have sealed their choice suggests readers are waiting this one out.
New Hampshire Senate
Slate readers say: Paul W. Hodes (D) 64 percent, Kelly Ayotte (R) 36 percent. Thirty-two percent of picks are locked.
Pundits say:Likely Republican (Rothenberg); leans Republican (Sabato, CQ); tossup (Cook).
This race is where Lean/Lock players disagree most with the pros. Silver's algorithms give Ayotte, a former state attorney general, a 91 percent chance of beating Hodes, who has represented the state in the U.S. House for two terms. But Lean/Lock players are betting otherwise: Even among the most confident players—those who are locking—Hodes has a 8 percentage-point edge.