Prediction Map

* = Pickup via defeat of incumbent; ^ = Pickup of an open seat

Prediction States Won

		19 |	37 |	37 |
Dem	10
Rep	27
Ind	0
Non	0

Confidence States Won

		19 |	37 |	37 |
Dem	9
Rep	26
Ind	0
Tos	2
Non	0

State Pick-ups

	Gain			Loss			Hold			Net Gain
	Inc.	Open	Total	Inc.	Open	Total	Inc.	Open	Total
Dem	0	0	0	-5	-4	-9	7	3	10	-9
Rep	+5	+4	+9	0	0	0	10	8	18	+9
Ind	0	0	0	0	0	0	0	0	0	0

Predicted Senate Control (112th Congress):

Party Seats Up Seats Not Up Total Seats Democratic 10 38 48 Republican 27 23 50 Independent 0 2 2

Prediction Score (max Score = 74)

Score	State Wins	State Percentages
55	33	22

Analysis

Last day prediction. I may update it later in the day, but here are my final calls (for now):
Wisconsin: Johnson +9.6
Alaska: Miller +6.2
Pennsylvania: Toomey +6.1
Illinois: Kirk +5.2
Nevada: Angle +4.2
Colorado: Buck +3.4
Washington: Rossi +1.4
West Virginia: Manchin +1.8
California: Boxer +4.8
Connecticut: Blumenthal +8.7
NPV: GOP +7.9
Net Change: GOP +8.8 seats
GOP Takeover Probability: 20%

Prediction History

Comments History - hide

Version: 86

New predictions:
Alaska: Miller +6.2 >
Pennsylvania: Toomey +5.8 >
Nevada: Angle +4.6
Illinois: Kirk +4.2
Colorado: Buck +2.9 >
Washington: Rossi +0.1
West Virginia: Manchin +1.4 >
California: Boxer +5.4 <

Version: 85

Washignton: Rossi +0.1
Pennsylvania: Toomey +3.5
NPV: GOP +6.6

My Senate takeover chances have gone up to 11%. I am most uncertain about the Washington prediction, since it is so close and the King County Elections Department may find an extra 1,500 ballots somewhere.

Version: 84

Predictions:
Pennsylvania: Toomey +3.6
Washington: Murray +1.5
California: Boxer +5.3

Version: 83

Predictions in important states:
Nevada: Angle +4.6
Illinois: Kirk +4.2
Pennsylvania: Toomey +3.5
Alaska: Miller +3.3
Colorado: Buck +2.2
West Virginia: Manchin +1.7
Washington: Murray +2.7
California: Boxer +5.0

Version: 81

Illinois: Kirk +3.9 >
Pennsylvania: Toomey +3.4 >
Colorado: Buck +2.3 >
West Virginia: Raese +0.8 <
Connecticut: Blumenthal +12.2 <

Version: 80

I have implemeted a better time frame for polls; instead of seeingif they are within a month of today's date, I will see if they were taken before the most recent gap in polling, if that exists. This has made the results much more reflective of the current state of the race:
Missouri: Blunt +9.6 <
Kentucky: Paul +5.6 >
Illinois: Kirk +3.8 >
Colorado: Buck +2.4 <
Pennsylvania: Toomey +3.1 <
California: Boxer +1.5 >
Washington: Murray +2.7 <

Additional statistics:
National Popular Vote: GOP +6.8
Chance of Winning Senate: 17%
Projected Seat Pickup: GOP +8.4
Likelyhood of error: 86%
Projected # of Errors: 1.7 states

Version: 79

New predictions in close states:
Pennsylvania: Toomey +3.8 <
California: Boxer +1.9 >
Washington: Murray +2.1 >
NPV: GOP +6.8

Version: 78

Updated predictions in competitive states:
Pennsylvania: Toomey +6.1 <
Kentucky: Paul +4.4 <
Illinois: Kirk +3.4 >
West Virginia: Raese +1.9 >
Washington: Murray +2.1 >
California: Boxer +2.5 >

Washington's mail ballot will make it difficult for Rossi to win, since he will not only have to pass Murray, but obtain enough of a lead so as to compensate for Murray's lead among people who have already voted. It's also difficult to see Fiorina winning, but it isn't impossible. Illinois, Nevada, and West Virginia have all turned toward the Republicans, while Pennsylvania, Kentucky, and Colorado are on the verge of being tossups.

Version: 77

Among the more notable changes:
Florida: Rubio +13.8 <
Missouri: Blunt +9.3 >
Wisconsin: Johnson +7.3 <
Pennsylvania: Toomey +6.7 <
Kentucky: Paul +6.0 <
Colorado: Buck +4.0 <
Nevada: Angle +2.0 >
West Virginia: Raese +1.6 <
Washington: Murray +3.0 >
California: Boxer +3.4 >

All of the changes in individual states in favor of the Democrats have dropped the GOP's generic ballot lead to 7.0

Version: 76

California: Boxer +3.4 >
Connecticut: Blumenthal +8.5 >
Ohio: Portman +21.3 >
NPV: GOP +7.8 >

The large shift in National Senate vote may bode well for GOP chances in the tossup states.

Version: 75

Alaska: Miler +3.3 <
Nevada: Angle +1.0 <

I have no idea what is happening in Alaska, so even though Murkowski will probably lose since she isn't on the ballot, to tossup status it goes.

Version: 74

The four tossup states:
Washington: Murray +1.3
West Virginia: Raese +1.9
Nevada: Angle +2.0
Illinois: Kirk +2.3

While GOP's national Senate lead remains steady at 7.2%, their takeover chances have fallen to 17%.

Version: 73

Washington: Murray +0.07 >
Illinois: Kirk +2.5 <
Nevada: Angle +2.3 <
West Virginia: Raese +3.9 <
Wisconsin: Johnson +10.4 <
Connecticut: Blumenthal +10.2 >
Delaware: Coons +14.6 <
Arizona: McCain +29.2 >
Florida: Rubio +19.3 >
Alaska: Miller +11.6 <

Based on momentum of the races, I would give Rossi and Giannoulias better chances than the model gives them.

Version: 72

Due to the closeness of the election, I've redefined the Lean category from being <10% to <6.8% (two standard deviations). In the two most important races, Angle leads by 2.4 and Murray leads by 1.0, but the momentum clearly favors Angle and Rossi. The GOP has a 27% chance of winning the Senate.

Version: 71

Changes of significance:
Nevada: Angle +2.4
Washington: Murray +0.8
Connecticut: Blumenthal +10.5
New York: Gillibrand +8.4

Looks like the summer of Harry and Charlie is finally over, and the Senate may be teetering on the edge as well.

Version: 70

Looks like Lincoln has jumped up to Santorum territory:
Arkansas: Boozman +20.3 <
California: Boxer +4.7 >
Colorado: Buck +7.9 >
Connecticut: Blumenthal +6.5 <
Illinois: Kirk +4.0 >
Maryland: Mikulski +24.8 >

Version: 69

Nevada now has company in the tossup category.
Washington: Murray +1.8
Kentucky: Paul +7.7
California: Boxer +5.6
Wisconsin: Johnson +12.3
Florida: Rubio +13.2
Illinois: Kirk +4.2
Colorado: Buck +3.6
Projected Seat Change: GOP +8.6
NPV: GOP +7.1 (this isn't a swing toward the Democrats; I had a computing error in my intial calculation)

Version: 67

UPDATES:
West Virginia: Raese +4.5
Connecticut: Blumenthal +4.1
Washington: Murray +4.1
New York: Gillibrand +8.3
Ohio: Portman +12.0
Pennsylvania: Toomey +8.3
Wisconsin: Johnson +10.4
Colorado: Buck +2.4
Illinois: Kirk +2.8
Georgia: Isakson +24.0
New York: Schumer +23.9
Nevada: Reid +0.2

NPV: GOP +8.9
Projected Seat Change: GOP +8.3 seats

It seems like Connecticut and West Virginia ought to be closer than they are. Only a slight increase in the GOP's projected national margin corresponded with picking up half a seat (most of that coming from West Virginia).

Version: 66

The only change of significance here is Kentucky, which is now Paul +8.9. In addition to plugging in recent polls, I have also devised a method for computing National Popular Vote in the Senate elections. Right now, the Republicans are leading the National Popular Vote by 8.4 points. Given that this is a national election, changes in the NPV will have effects on individual Senate races. If the GOP gets past 10 points, they will probably win Colorado, Illinois, Nevada, and West Virginia, and move California, Connecticut, New York, and Washington into the tossup category. If the GOP advantage slips below 5, they probably won't win Colorado, Illinois, Nevada, or West Virginia, and they might be vulnerable in Missouri, Kentucky, or New Hampshire. Interestingly, if NPV were GOP +0.0, the Republicans would still be favored to pick up a net 3 seats.

UPDATES:
New Hampshire: Ayotte +10.0
Kentucky: Paul +8.9
Connecticut: Blumenthal +5.6
Pennsylvania: Toomey +10.0
Ohio: Portman +11.8
California: Boxer +5.2
Florida: Rubio +12.3
Projected Seat Change: GOP +7.8
National Popular Vote: GOP +8.6

Version: 65

New predictions:
California: Boxer +5.0
Washington: Murray +3.9
New York: Gillibrand +8.8
New York: Schumer +24.3
Missouri: Blunt +8.1
Nevada: Angle +0.3
North Carolina: Burr +19.4
Projected GOP gain: +7.7 seats

GOP gains in West Virginia, Wisconsin, and New York were offset by Democratic gains in California and Washington.

Version: 64

New predictions:
California: Boxer +2.3
Wisconsin: Johnson +10.1
Pennsylvania: Toomey +10.1
Illinois: Kirk +2.0
Colorado: Buck +2.4
Arkansas: Boozman +32.0
Delaware: Coons +14.5
Projected GOP Gain: +8.0 seats

Version: 63

New predictions:
West Virginia: Manchin +0.7
Wisconsin: Johnson +7.5
California: Boxer +0.8
Alaska: Miller +18.0 (over Murkowski)
Delaware: Coons +14.3
Pennsylvania: Toomey +10.2
Ohio: Portman +11.8
New York: Gillibrand +15.4

Projected GOP pickup: +8.3 seats

Version: 62

California: Boxer +1.0
West Virginia: Manchin +7.8

Murkowski's write-in candidacy creates uncertainty with my prediction, so for now, I am predicting Miller 48-McAdams 36-Murkowski 15-Haase 1

California's flip returns this prediction to the mean, since the average Republican pickup is 7.6 seats.

Version: 61

New predictions:
Wisconsin: Johnson +2.4
Illinois: Kirk +2.4
Ohio: Portman +10.8
Indiana: Coats +19.2
Oregon: Wyden +19.1
Washington: Murray +1.2
New Hampshire: Ayotte +8.0
Arkansas: Boozman +33.8
Nevada: Angle +0.2
California: Fiorina +0.9

Version: 60

New predictions:
Florida: Rubio +11.8
Ohio: Portman +11.2
Pennsylvania: Toomey +10.8
Delaware: Coons +14.2
New Hampshire: Ayotte +10.3
Nevada: Tied

Nevada came out of my prediction as a perfect tie, so I just nudged my Atlas prediction toward Angle by a hair partly out of hackishness and partly because she seems to have a bit of momentum. While the GOP might be slight favorites in enough contests to take the Senate as of now, on average they take only 7.8 seats when the probabilities of Democratic wins are factored in.

Version: 59

New margins:
Connecticut: Blumenthal +8.0
Nevada: Reid +0.3
Ohio: Portman +10.2
Washington: Rossi +1.1
North Carolina: Burr +16.3
Kentucky: Paul +12.8
Georgia: Isakson +18.9

Version: 58

And the Democrats take back their theoretical majority. Republican chances of winning nominal control are reduced from ~36% to ~10%:
Delaware: Coons +8.4
Nevada: Reid +1.8

Version: 57

Delaware and New Hampshire are provisionally moved to tossup until the primaries occur.
Other states:
Connecticut: Blumenthal +9.8
North Carolina: Burr +12.2
Illinois: Kirk +3.0
Florida: Rubio +8.9
West Virginia: Manchin +5.4
Missouri: Blunt +9.0
Arizona: McCain +16.9
California: Fiorina +0.7
Kentucky: Paul +13.5

Version: 56

For the first time in this cycle, I am projecting a GOP Senate majority: (partly due to the Rasmussen poll, partly from discarding old polls)
Califoria: Fiorina +2.9

Version: 55

Kentucky: Paul +12.3
Ohio: Portman +9.5

Version: 54

Missouri: Blunt +10.1
Pennsylvania: Toomey +10.1
North Carolina: Burr +10.1
Alaska: Miller +9.9
Ohio: Portman +8.5
Kentucky: Paul +7.9
Florida: Rubio +7.6
Washington: Rossi +4.5
Colorado: Buck +3.8
Illinois: Kirk +1.4
Wisconsin: Johnson +0.2
Nevada: Reid +1.6
California: Boxer +3.2
West Virginia: Manchin +6.5

Version: 53

Colorado: Buck +4.1
Pennsylvania: Toomey +9.9

Version: 52

PPP's new poll makes Rubio the clear favorite:
Florida: Rubio +5.0
Missouri: Blunt +11.0
Illinois: Kirk +2.5
West Virginia: Manchin +19.8
Oregon: Wyden +21.4

I now give the GOP a 27% chance of winning the Senate.

Version: 51

Looks like I spoke too soon about Washington:
Washington: Murray +0.4
Georgia: Isakson +14.4
Arkansas: Boozman +34.9
Louisiana: Vitter +16.9
Kentucky: Paul +7.1

Based on my simulations, I would estimate that the GOP has a ~9% chance of taking the Senate as of today.

Version: 50

Due to the volume of polling coming out of Florida, I limited the time frame to be no more than one month old. Thus, I imputed the new Quinnipiac poll and discarded the old PPP poll:

Florida: Rubio +1.4

Likewise, I discarded the three-month old Washington poll, which probably had a greater effect on Murray's margin than the Rasmussen poll did. Additionally, the fact that Murray increased her margin of victory over Rossi after the primary doesn't bode well for Rossi's chances:

Washington: Murray +3.8

The GOP is now favored to pick up between 7 and 8 senate seats.

Version: 49

Given that Realclearpolitics.com and Pollster.com have both included Rasmussen's leaners, I will do the same:
Pennsylvania: Toomey +7.8
Kentucky: Paul +8.5
Ohio: Portman +4.8
Nevada: Angle +0.2

I am currently seeing how adding Washington's jungle primary results into my data affects my model. If I can create a model with a lower standard error than my current best one (+/- 3.4%) then I will start using that as my main model.

Version: 48

New predictions:
Ohio: Portman +4.4
Pennsylvania: Toomey +8.2
Missouri: Blunt +8.6
Illinois: Kirk +0.5
Florida: Crist +0.3
North Dakota: Hoeven +52.1
Georgia: Isakson +16.9

My model now forecasts that it is even more likely that the Republicans gain 7 seats or more. They also now have a 4% chance at taking back the Senate. For this reason, I am tempted to adjust my Florida prediction, but I will wait until more updated polling comes out.

Version: 47

New predictions:
Florida: Crist +0.3
Connecticut: Blumenthal +8.6
Nevada: Reid +1.3
California: Boxer +2.3
Colorado: Buck +2.5
Missouri: Blunt +7.4

My simulations now show that a 7-seat gain for Republicans is the most likely outcome. Given that Wisconsin and Florida are essentially coin-flips at this point, I'm not particularly surprised.

Version: 46

New predictions:
Illinois: Kirk +1.1
Florida: Crist +1.9 (w/ Meek as DEM nominee)
Ohio: Portman +2.8
Colorado: Buck +2.9
Delaware: Castle +16.3
Iowa: Grassley +28.5
New Hampshire: Ayotte +10.1
Wisconsin: Feingold +0.3

Although currently if the favorite in each state wins, the Republicans will have a net gain of six seats, 300 simulations of the election show that a 7 seat gain is more likely, although not significantly.

Version: 45

Updates:
North Carolina: Burr +9.2
Kentucky: Paul +8.5
California: Boxer +3.2
Kansas: Moran +39.2

Version: 44

Several updates:
Kansas: Moran +42.7
Indiana: Coats +25.1
Arizona: McCain +22.8
North Carolina: Burr +11.0
Kentucky: Paul +6.8
Ohio: Portman +1.3
Nevada: Reid +1.2
Washington: Murray +2.2
Florida: Crist +3.0 (w/ Meek as DEM nominee)
Connecticut: Blumenthal +12.2
New York: Gillibrand +20.9

Version: 43

New predictions:
Wisconsin: Johnson +0.06
California: Boxer +3.7
Missouri: Blunt +6.8
New Hampshire: Ayotte +9.7
Oregon: Wyden +19.1

As of today, I can say with reasonable confidence that the GOP will pick up between 3 and 10 Senate seats. With less confidence, I can say that they will pick up between 5 and 8 Senate seats.

Version: 42

It may be too early to say, but it seems like the Republicans peaked too soon:
Illinois: Giannoulias +0.07
Nevada: Reid +2.6

Version: 41

New margins:
California: Boxer +3.8
New Hampshire: Ayotte +10.5
Colorado: Buck +2.9
North Carolina: Burr +11.9

Version: 40

The Mason-Dixon poll makes Missouri LEAN GOP:
Blunt +4.5

Version: 39

Discarding the outdating Survey USA poll explains Washington's switch:
Murray +0.08

Also, the Rasmussen and PPP polls now make it so Rubio is only 17 points ahead of Meek. However, this is still not enough to get Crist below 40% according to my model. If the election were today, I'd guess Crist 40-Rubio 37-Meek 20-Snitker 3

Version: 38

New margins:

Arkansas: Boozman +29.3
Pennsylvania: Toomey +3.9
Wisconsin: Feingold +0.3
Nevada: Reid +0.7
Florida: Crist +2.7
Connecticut: Blumenthal +15.6
West Virginia: Manchin +15.9

Version: 37

It seems that Harry's smear campaign is working.

Delaware: Castle +13.4
Washington: Rossi +4.5
Wisconsin: Feingold +0.8
Nevada: Reid +1.3
Connecticut: Blumenthal +21.5

Version: 36

Here's my new projections based on recent polls (taken from the prediction with lowest standard deviation):
Indiana: Coats +23.0
Colorado: Buck +4.1
Pennsylvania: Toomey +3.5
Missouri: Blunt +2.8
Nevada: Angle +2.1
Wisconsin: Johnson +1.1
Florida: Crist +2.1
California: Boxer +2.1
West Virginia: Manchin +15.9

Version: 35

Illinois and Wisconsin are the predictions I am most uncertain about, since those are the only states where my models disagree with each other. My models split 5-3 in favor of Ron Johnson (although 2 of the 5 are by 0.02%), and 7-1 in favor of Mark Kirk. While the prediction with the least standard deviation predicts Johnson and Kirk, two others that was closest in more states in 2008 and the 2010 Massachusetts special election predict Feingold and Kirk and Feingold and Giannoulias, respectively. The model that I had previously placed the most confidence in is now the least accurate after my updates. Here are what all eight models predict about those races:

Illinois: Kirk +0.65/Kirk +1.01/Kirk +0.18/Kirk +0.85/Kirk +0.23/Kirk +0.61/Giannoulias +0.20/Kirk +0.70

Wisconsin: Johnson +0.02/Johnson +0.38/Feingold +0.45/Johnson +0.26/Feingold +0.34/Johnson +0.03/Feingold +0.78/Johnson +0.12

Version: 34

Making more updates to my formula combined with new polls is the cause of the changes. Here's some important ones:
Colorado: Buck +2.3
Illinois: Kirk +0.6
Florida: Crist +1.3
California: Boxer +3.2

Version: 33

Rasmussen's new poll cancels out my update yesterday.
North Carolina: Burr +12.6 (NEW: Burr +13.0)

UPDATE #2: With Research 2000 being exposed and ostracized from the polling community, I have cleared their polls entirely from my projections. Doing this has reduced my standard error from 4.3% to 3.9% and increased by average error from 0.0% to -0.4% (GOP is positive, DEM is negative). Thus, I've adjusted my predictions to account for this.

Version: 32

Here are the changes:
Oklahoma: Coburn +41.0
North Carolina: Burr +7.4
Kentucky: Paul +5.9
Ohio: Portman +1.3

One hard aspect to factor into these predictions is what share of the vote third-parties will take. Illinois and North Carolina both have polls showing 3rd party candidates in double digits, but their vote shares will probably decline as the election nears. However, since Kirk and Burr both have decent leads, that decline probably won't be as substantial as it was in Chris Daggett's campaign last year. Thus, both will probably be held under 50%.

Version: 31

New polls in KY, MO, and OH change my projections:
Kentucky: Paul +6.2
Missouri: Blunt +5.6
Ohio: Portman +0.2

Version: 30

Important Changes:rnNorth Carolina: Burr +12.5rnKentucky: Paul +6.5rnPennsylvania: Toomey +4.5rnNevada: Angle +3.5rnWisconsin: Johnson +0.02rnrnWashington, Wisconsin, and Florida are my least confident states.

Update: I got rid of all polls 3 months or older, which may my outlook slightly more in tune with forecasters. Here's the NEW new changes:
Delaware: Castle +25.5
Pennsylvania: Toomey +3.7
Colorado: GOP +3.0
Washington: Rossi +2.2
Missouri: Blunt +1.1
Wisconsin: Johnson +0.02
Ohio: Fisher +0.3
Florida: Crist +3.0

Version: 29

Crist just barely has the advantage, with a 0.01% lead.

Version: 28

Changes:
Arkansas: Boozeman +25.6
Florida: Rubio +0.5
Kentucky: Paul +6.5
Illinois: Kirk +6.2
Nevada: Angle +5.0
Missouri: Blunt +4.9
Washington: Rossi +0.4
California: Boxer +3.9

Version: 26

Changes in Wisconsin and Connecticut can be attributed to scrapping old polls in light of recent events. Feingold's margin over Johnson is now 2.2, while Blumenthal's margin over McMahon is now 14.1 (I would like to include Schiff, but I am doubtful of his chances)

Other races:
Kentucky: Paul +12.9
Washington: Murray +1.3
California: Boxer +3.8

Version: 25

Explanation of changes:
Connecticut: The suddenness of Blumenthal's scandal hinders my model's ability to come up with an accurate prediction. As more polls come in I will be able to update it, but for now I suspect that Blumenthal has slumped into at least the 50s.

North Dakota: Updating the polls in my model shows Hoeven with a 54 point lead, which is enough to make it >70%

Washington: Rossi's expected entry lowers Murray's lead down to 2.9, enough to make it a tossup.

Version: 24

Sorry for the constant updates:

California: Boxer +0.8 over Campbell (+3.8 over Generic GOP)
Kentucky: Paul +7.0 over Conway
Arizona: McCain +22.7 over Glassman
Colorado: GOP +3.0 over DEM
Connecticut: Blumenthal +22.1 over GOP

Version: 23

Rubio is back to a +3.6 advantage over Crist.

Version: 22

Not to count my chickens before they hatch, but Campbell has a +0.3 lead over Boxer, and he unfortunately remains the primary frontrunner.

Version: 21

Idaho: Crapo +48.9
Kansas: GOP +36.6
Kentucky: Paul +9.1
Pennsylvania: Toomey +6.4
California: Boxer +2.7

Version: 20

Of interest:
Arkansas: Boozeman +22.9
Iowa: Grassley +12.4
Illinois: Kirk +9.1
Pennsylvania: Toomey +7.1 (over Sestak)
Ohio: Portman +1.8
Florida: Crist +0.8

Version: 19

New margins:
Louisiana: Vitter +18.2
Iowa: Grassley +16.4
Missouri: Blunt +8.8
Florida: Crist +0.2 (over Rubio)

Version: 18

New margins:
Arizona: McCain +20.9
Delaware: Castle +20.9
Georgia: Isakson +20.1
Arkansas: GOP +14.2
New Hampshire: Ayotte +13.2
Illinois: Kirk +5.6
Ohio: Portman +3.2

Version: 17

Important Changes:
FL: Crist is running as an independent, so Rubio's margin against Crist slips to 3.9%. Due to the three-way nature of the race, this may be the most diffucult to predict.
WA: Murray's margin against Benton is down to 7.1%.
SD: Thune is unopposed, so his margin goes up accordingly.

Version: 16

New margins:

Kentucky: Paul +12.1
Nevada: GOP +11.9
New Hampshire: Ayotte +10.9
Missouri: Blunt +7.7
Pennsylvania: Toomey +5.5
Ohio: Portman +4.2
Illinois: Kirk +4.1
Colorado: GOP +2.7
Washington: Murray +8.9 (against Benton)
Wisconsin: Feingold +11.0 (against generic GOP)

Version: 15

Changes:
Ohio Portman +4.6
Missouri Blunt +4.1
California Boxer +0.02 (against Campbell)
Boxer +2.7 (against generic opponent)
Illinois Giannoilias +3.6
Pennsylvania Toomey +3.6 (against Specter)
Colorado GOP +1.7

Version: 14

If Pennsylvania and California are adjusted so that Campbell and Specter are the general election candidates, Toomey's advantage falls to 4.2% and Boxer's advantage falls to 2.7%.

Version: 12

The recent Kos poll, while of dubious value, shifts the Illinois race from 0.1% Kirk to 0.2% Alexi

Version: 11

Bayh's retirement gives yet another seat to the GOP. Hopefully Hostettler will beat Coats.

Version: 10

The new Rasmussen poll flips Illinois, which I now project Kirk will win by .13%

Version: 9

With new polls from PPP & Rasmussen, my formula is now putting Arkansas at GOP +10.0, allowing me to put it as LEAN GOP.

Version: 8

Feingold and Gillibrand are both in the category of safe unless a certain person runs, in which case it's a tossup. I put them as lean for good measure.

Version: 7

The new polls out in Pennsylvania put Toomey's projected lead up to 9.8%, which I consider lean.

For some reason, I am getting closer results for Arkansas than one would expect, with Lincoln's deficit at a mere 1.7%, so still a tossup.

Additionally, Biden's withdrawal leaves me with an obsolete projection for Delaware, so I can't make an accurate projection there until new polling comes out. Until then, I doubt Castle will have a tough race.

UPDATE: For races that I have little data on, I have come up with a formula that roughly correlates my formula's spread with the Intrade.com spread. Thus, I have Castle winning by about 13.9%, which I would consider strong. Of course one must keep in mind that this will have a greater MoE than the original formula.

Version: 6

Glad to see I pegged Bayh as more vulnerable than he appeared before the Rasmussen poll came out.

Version: 5

I was amazed that the formula that I am using for this Senate election cycle was able to get the Massachusetts senate election result correct to within a tenth of a percentage point.

Version: 4

Also, Massachusetts is R pickup >50% Tossup

Version History

87: 2010-11-01 @ 17:25:39 86: 2010-10-31 @ 23:36:05 85: 2010-10-29 @ 18:23:37 84: 2010-10-28 @ 18:58:28 83: 2010-10-27 @ 18:12:23 82: 2010-10-26 @ 18:04:21 81: 2010-10-25 @ 20:44:38 80: 2010-10-23 @ 22:38:54 79: 2010-10-22 @ 19:09:23 78: 2010-10-20 @ 18:07:26 77: 2010-10-19 @ 18:34:36 76: 2010-10-16 @ 02:21:12 75: 2010-10-14 @ 18:25:43 74: 2010-10-13 @ 17:58:29 73: 2010-10-12 @ 18:54:15 72: 2010-10-09 @ 11:40:01 71: 2010-10-07 @ 17:59:09 70: 2010-10-04 @ 17:55:17 69: 2010-09-30 @ 17:56:05 68: 2010-09-29 @ 18:23:54 67: 2010-09-28 @ 19:12:30 66: 2010-09-26 @ 23:53:26 65: 2010-09-23 @ 18:26:28 64: 2010-09-22 @ 18:31:04 63: 2010-09-21 @ 18:08:21 62: 2010-09-20 @ 18:05:32 61: 2010-09-17 @ 17:59:44 60: 2010-09-15 @ 19:17:17 59: 2010-09-14 @ 18:44:17 58: 2010-09-13 @ 00:08:26 57: 2010-09-12 @ 03:16:34 56: 2010-09-07 @ 18:48:28 55: 2010-09-05 @ 15:23:44 54: 2010-09-04 @ 03:39:43 53: 2010-08-25 @ 13:41:34 52: 2010-08-24 @ 13:23:16 51: 2010-08-21 @ 12:47:21 50: 2010-08-19 @ 14:43:48 49: 2010-08-18 @ 18:35:34 48: 2010-08-17 @ 13:26:11 47: 2010-08-15 @ 03:12:41 46: 2010-08-11 @ 18:49:28 45: 2010-08-06 @ 12:50:08 44: 2010-08-04 @ 14:50:55 43: 2010-07-29 @ 19:16:58 42: 2010-07-28 @ 13:27:04 41: 2010-07-27 @ 16:16:34 40: 2010-07-24 @ 13:41:06 39: 2010-07-22 @ 18:08:05 38: 2010-07-20 @ 14:31:34 37: 2010-07-16 @ 11:51:26 36: 2010-07-15 @ 17:17:06 35: 2010-07-11 @ 23:24:17 34: 2010-07-09 @ 02:20:33 33: 2010-07-07 @ 00:03:34 32: 2010-07-06 @ 18:37:54 31: 2010-06-30 @ 22:46:06 30: 2010-06-29 @ 20:09:23 29: 2010-06-21 @ 18:36:52 28: 2010-06-16 @ 19:01:00 27: 2010-06-09 @ 19:57:53 26: 2010-05-27 @ 19:48:27 25: 2010-05-24 @ 19:08:14 24: 2010-05-20 @ 00:31:58 23: 2010-05-17 @ 22:12:19 22: 2010-05-15 @ 13:12:14 21: 2010-05-14 @ 19:29:40 20: 2010-05-11 @ 00:35:34 19: 2010-05-04 @ 20:01:22 18: 2010-05-01 @ 01:53:56 17: 2010-04-28 @ 19:22:54 16: 2010-04-10 @ 03:06:26 15: 2010-03-30 @ 19:59:37 14: 2010-03-16 @ 19:21:53 13: 2010-03-15 @ 00:44:46 12: 2010-02-28 @ 23:40:15 11: 2010-02-21 @ 00:33:04 10: 2010-02-04 @ 17:33:12 9: 2010-02-02 @ 16:48:25 8: 2010-01-28 @ 18:46:49 7: 2010-01-27 @ 19:06:04 6: 2010-01-25 @ 18:59:58 5: 2010-01-22 @ 00:15:03 4: 2010-01-15 @ 19:16:55 3: 2010-01-10 @ 23:48:33 2: 2010-01-02 @ 23:55:44 1: 2009-11-21 @ 10:39:26 View Version