Nate Silver has unveiled his algorithms for rating the chances of each party in this year’s Senate races. This replaces his seat-of-the-pants ratings which seemed to me to be unduly optimistic for his side (the Democrats). Silver’s numbers attempt to show the percentage likelihood of each party winning the seat, taking into account poll results, candidates’ ratings and the likelihood of candidates to win their parties’ nominations.
Silver’s bottom line is the same as that of Larry Sabato: at the moment, Democrats are likely to lose seven seats and pick up none. That would reduce the Democrats’ majority in the Senate to 52-48. But one should note that Silver sees the likelihoods of the Illinois seat going Republican at only 51%. Here are Silver’s percentage likelihoods of party changes for Republican-held and Democratic-held seats:
Republican seats Democratic seats
Missouri 37 North Dakota 99
New Hampshire 35 Arkansas 73
Ohio 35 Nevada 73
Kentucky 26 Pennsylvania 72
Florida 18 Colorado 70
North Carolina 15 Delaware 66
Louisiana 7 Illinois 51
Georgia 5 Indiana 37
Arizona 3 California 21
Iowa 3 Wisconsin 16
Kansas 1 New York (short term) 13
Alabama 0 Washington 7
Alaska 0 Hawaii 4
Idaho 0 Oregon 3
Oklahoma 0 Connecticut 1
South Carolina 0 Maryland 0
South Dakota 0 New York (long term) 0
Utah 0 Vermont 0
My responses: The 72% likelihood of party change in Pennsylvania feels high to me, and the 1% likelihood in Connecticut feels low. (Didn’t a seemingly popular Democratic state attorney general just lose in Massachusetts?)
I note that if such estimates had been made a year ago today, the likelihood of party change in three or four Republican seats (MO, NH, OH and maybe KY) would have been over 50%, and the likelihood of party change in four Democratic seats would have been zero or close to that (IN, CA, WI and NY short term).
