Top Senate predictions

(Great stuff.   – promoted by Garrett Quinn)

Based on the total deviations method (thank you to BMG’s Trickle Up) taken from the data on the predictions aggregate page, here are the top 5 RMG predictions for the January 19th Senate race:









































































Deviations Predictor Brown Coakley Kennedy
0 Actual 51.9 47.1 1





0.2 Ken Pittman 52 47 1
0.2 Paul “Cool Cal“ Ferraro 52 47 1
1.6 Mike “DD4RP” Rossettie 52 46.2 1.6
1.8 yankeepundit 51 48 1
1.8 Republican Ram Rod Radio 51 48 1
2.5 RMG prediction average 51 46.9 2.4
10.3 BMG prediction average 46.7 50.4 2.8

Congrats to Ken Pittman and Paul “Cool Cal” Ferraro. You really couldn’t have gotten it any more dead on. Overall, RedMass got these predictions really close, and the RedMass average prediction was closer than any single BlueMass prediction.

Full listing (including BMG predictions in blue) below the fold, along with some possible reasons for the discrepancy.




































































































































































































































































































































































Deviations Predictor Brown Coakley Kennedy
0 Actual 51.9 47.1 1





0.2 Ken Pittman 52 47 1
0.2 Paul “Cool Cal“ Ferraro 52 47 1
1.6 Mike “DD4RP” Rossettie 52 46.2 1.6
1.8 yankeepundit 51 48 1
1.8 Republican Ram Rod Radio 51 48 1
2 garyrlake 51 47 2
2 swamp_yankee 51 47 2
2.5 RMG prediction average 51 46.9 2.4
2.8 Rob “EaBo Clipper” Eno 50.5 48.5 1
3.8 Argyle 50 49 1
4 South Shore Republican 50 47 3
4.2 demolisher 54 45 1
5 JoeTS 49 47 3
5 Peter Porcupine 49 47 3
5.6 Philo Publius 53.7 44.3 2
5.8 JohnD 49 48.5 2.5
5.8 MerrimackMan 49 48 3
6 cater68 51 45 4
6 Peter Porcupine 49 47 4
6 All predictions average 48.9 48.7 2.4
6.1 BlueMass model 52.7 44 3.2
6.2 VESuggestions 54.3 44 1.7
6.2 Witch7 54 44 2
6.2 nomad943 54 44 2
7.6 ElectricStrawberry 48.1 48.1 3.8
7.8 sleeples 48 50 2
7.8 sleepleswithamoustache 48 50 2
7.8 Vote3rdpartynow 48 48 4
8.1 BMG/RMG models AVG 47.9 49.2 3
8.8 Hoyapaul 47.5 51 1.5
8.8 jconway 47.5 49.5 3
8.8 PeteSimon 47 49 3
9.2 Manny Happy Returns 53 41 3
9.8 sabutai 47 51 2
10.3 BMG prediction average 46.7 50.4 2.8
11.8 Wookie 46 52 2
11.8 HLPeary 46 52 2
11.8 BigD 46 49 5
12.8 ruppert 46 53 2
13.8 Carey Theil 45 53 2
13.8 Ray M 45 52 3
15.8 eddiecoyle 44 54 2
15.8 PatrickLong 44 51 5
16.8 Steve Stein 43 51 5
19.6 RedMass model 42.1 55.2 2.7
21.8 BrooklineTom 46 61 3
32.8 petr 32 59

Possible reasons for why RedMass did so well?

1) In a tossup race, BMG erred on the side of Coakley, RMG on the side of Brown. Race broke to Brown, RMG got lucky.

2) RMG better at predicting outcomes of elections.

or 3) (based on feedback from here) With nothing actually riding on the predictions, each candidate’s supporters wants to motivate their base by saying their own candidate will win in a squeaker if they just GOTV enough.

Future data may help answer that question.

About sleepleswithamoustache