Past as predictor of future: a Senatorial psycohistory

                                     Hari Seldon ponders the Senate race
On the one hand you could just check the pollster.com Senate charts to guess the winners and losers (although GA is going to be a Dem win) or you can play a history-based model out and see what things look like.

Simply put, close Senate races tend to break mostly all one way or the other. The important thing is that there are always one or two exceptions. But whether there is one exception this year or two exceptions is the likely difference between the Dems achieving their filibuster-proof majority dream or not.

This year, the Dems have pick-up opportunities in 10 seats: VA, NM, CO, AK, NH, MN, NC, GA, MS and KY.

In 2006, the Dem wave carried 6 out of 7 Senate targets (VA, MO, MT, OH, PA, RI but failed in TN) - a near 85% success rate. In 2004, the GOP wave carried 6 out of 8 target seats (FL, GA, LA, NC, SD, SC but failed in IL and CO) - a 75% success rate. In 2002, the GOP wave carried 3 out of 4 target seats (GA, MN, MO but failed in AR) - a 75% success rate.

Thus the average success rate was 78.3%. Applying that to this year's Senate seats, it would seem that Dems are on course to pick up 8 out of 10 seats, likely losing only KY and MS - per those pollster.com averages!