Election simulation

[Inverted Things]

I wrote a C program to simulate the Tuesday election. Here are my assumptions.

Safe Bush states (227):
Alabama (9)
Alaska (3)
Arizona (10)
Arkansas (6)
Colorado (9)
Georgia (15)
Idaho (4)
Indiana (11)
Kansas (6)
Kentucky (
Louisiana (9)
Mississippi (6)
Missouri (11)
Montana (3)
Nebraska (5)
Nevada (5)
North Carolina (15)
North Dakota (3)
Oklahoma (7)
South Carolina (
South Dakota (3)
Tennessee (11)
Texas (34)
Utah (5)
Virginia (13)
West Virginia (5)
Wyoming (3)

Safe Kerry states (206):
California (55)
Connecticut (7)
Delaware (3)
D.C. (3)
Hawaii (4)
Illinois (21)
Maine CD1 + State (3)
Maryland (10)
Massachusetts (12)
Michigan (17)
New Jersey (15)
New York (31)
Oregon (7)
Rhode Island (4)
Vermont (3)
Washington (11)

Swing States:
Maine CD2 (1)
New Hampshire (4)
New Mexico (5)
Iowa (7)
Minnesota (10)
Wisconsin (10)
Ohio (20)
Pennsylvania (21)
Florida (27)

Then I assigned probabilities of Kerry winning these states (I've taken into account everything to date, including the Mason-Dixon poll dump).
ME CD2--90%
NH--80%
NM--20%
IA--30%
MN--50%
WI--50%
OH--45%
PA--85%
FL--25%

I've assumed that the states are independent of each other. The flaw in this assumption is that I end up assuming Kerry has the same chances of winning FL regardless of whether or not he gets PA, which is probably not true.

I ran this as a Monte Carlo simulation 1,000,000 times and concluded that Kerry wins 22% of the time, with a 5% chance of a 269-269 tie.

Comments? Suggestions?

[© tweed]

ME CD2--90%
NH--80%
NM--20%
IA--30%
MN--50%
WI--50%
OH--45%
PA--85%
FL--25%

ME CD2--80%
NH--60%
NM--30%
IA--35%
MN--50%
WI--40%
OH--45%
PA--75%
FL--45%

[Inverted Things]

Boss-- Your values give Kerry a 30% chance, with a 3.5% chance for a tie.

[© tweed]

Boss-- Your values give Kerry a 30% chance, with a 3.5% chance for a tie.

You can calculate it that quickly?

[Inverted Things]

Yup, on a 300 MHz processor, no less.

[opebo]

Ok how about this -

ME CD2--90%
NH--80%
NM--40%
IA--40%
MN--52%
WI--50%
OH--55%
PA--85%
FL--49%

[Inverted Things]

Opebo, your values give Kerry a 44.4% chance, with a 4.7% chance for a tie

[Umengus]

and...

Me CD2: 75%
NH: 75%
NM: 30%
IA: 53%
MN: 55%
WI: 55%
OH: 50%
PA: 79%
FL: 48%

[Inverted Things]

Umengus, Kerry gets a 42.0% chance, 4.2% chance for a tie

[Umengus]

Umengus, Kerry gets a 42.0% chance, 4.2% chance for a tie

thanks!

[Angel of Death]

Since you're running a Monte Carlo simulation, couldn't you include all the states to account for the odd surprise?
Perhaps someone can calculate the chances more systematic based on a combination of state polls.

[CARLHAYDEN]

I must disagree with some of your "assumptions," with explanations.

First, ballot measures on several states are impacting the vote there:

Maine          In addition to the 'hunters rights' ballot measure there is
                   'Prop 13" style property tax limitation on the ballot.  Kerry
                   is NOT assured of winning the state!

Michigan      The 'marriage' measure on the ballot combined with a late
                   (but effective) bush campaign there pointing out Kerry's
                   support for Kyoto and increased CAFE standards has
                   also made this a swing state.

Oregon       Oregon has two measures on the ballot which help Bush.
                   First is the 'marriage' measure and second is the medical
                   malpractice tort reform measure.  This is also a swing
                   state.

Second, I disagree with your assignment of probabilities on the states you listed as 'swing' states.  You are far to generous in several of them to Kerry.  I'll only list those that are significantly off the mark.

Specifically,

Maine CD2          45% rather than 90% would be more accurate.
New Hampshire  45% would be more accurate than 80%.
Ohio                    45% would be more accurate than 55%.
Pennsylvania       52% would be more accurate than 85%

[kevinatcausa]

I think the net effect of assuming independence will always be to skew the results away from 50-50 odds, regardless of who is ahead (the actual results will tend to have more blowouts than your simulation, so a larger standard deviation while keeping the same expected value).  I don't know how to adjust for that, but it's just something to keep in mind.

[Angel of Death]

Don't make the classic mistake of equating the expected share of the vote with the chance of winning a plurality of votes. Pennsylvania is much more likely to go to Kerry than 52%. BTW, where is Nebraska (and its districts)?

[Inverted Things]

Since you're running a Monte Carlo simulation, couldn't you include all the states to account for the odd surprise?
Perhaps someone can calculate the chances more systematic based on a combination of state polls.

I could, but in my opinion that would give less than reliable results. Among my biggest assumptions is independence of states. If I give Bush a chance at Washington, for instance, Bush likely has Florida all locked up, which would violate my key assumption of state independence.  I might add CO, WV, NV, MI, OR, NJ and ME at large, but I feel a little sketchy about those as well.

[Umengus]

I must disagree with some of your "assumptions," with explanations.

First, ballot measures on several states are impacting the vote there:

Maine          In addition to the 'hunters rights' ballot measure there is
                   'Prop 13" style property tax limitation on the ballot.  Kerry
                   is NOT assured of winning the state!

Michigan      The 'marriage' measure on the ballot combined with a late
                   (but effective) bush campaign there pointing out Kerry's
                   support for Kyoto and increased CAFE standards has
                   also made this a swing state.

Oregon       Oregon has two measures on the ballot which help Bush.
                   First is the 'marriage' measure and second is the medical
                   malpractice tort reform measure.  This is also a swing
                   state.

Second, I disagree with your assignment of probabilities on the states you listed as 'swing' states.  You are far to generous in several of them to Kerry.  I'll only list those that are significantly off the mark.

Specifically,

Maine CD2          45% rather than 90% would be more accurate.
New Hampshire  45% would be more accurate than 80%.
Ohio                    45% would be more accurate than 55%.
Pennsylvania       52% would be more accurate than 85%

I tought that Kerry was against kyoto.

[CARLHAYDEN]

Kerry opposes the Kyoto treaty per se, but favors the measures in the Kyoto treaty for the United States.

He would support enacting the measures in the Kyoto treaty even if it other nations do not adopt the same measures.

So, substantively Kyoto favors the Kyot treaty while saying he opposes it.

[Inverted Things]

I must disagree with some of your "assumptions," with explanations.

First, ballot measures on several states are impacting the vote there:

Maine          In addition to the 'hunters rights' ballot measure there is
                   'Prop 13" style property tax limitation on the ballot.  Kerry
                   is NOT assured of winning the state!

Michigan      The 'marriage' measure on the ballot combined with a late
                   (but effective) bush campaign there pointing out Kerry's
                   support for Kyoto and increased CAFE standards has
                   also made this a swing state.

Oregon       Oregon has two measures on the ballot which help Bush.
                   First is the 'marriage' measure and second is the medical
                   malpractice tort reform measure.  This is also a swing
                   state.

Second, I disagree with your assignment of probabilities on the states you listed as 'swing' states.  You are far to generous in several of them to Kerry.  I'll only list those that are significantly off the mark.

Specifically,

Maine CD2          45% rather than 90% would be more accurate.
New Hampshire  45% would be more accurate than 80%.
Ohio                    45% would be more accurate than 55%.
Pennsylvania       52% would be more accurate than 85%

With regard to Maine, I've seen one poll showing Bush ahead. That was in mid September when Bush was at high tide.

Michigan- Most recent poll showing a Bush lead was in June. Not my kind of swing state.

Oregon-- One outlier poll in October showing Bush +5. Rest of October polls show Kerry +1-Kerry+9. Again, not my kind of swing state.

With your values, Kerry has a 14.9% chance with a 1.3% chance to tie

[Inverted Things]

Don't make the classic mistake of equating the expected share of the vote with the chance of winning a plurality of votes. Pennsylvania is much more likely to go to Kerry than 52%. BTW, where is Nebraska (and its districts)?

Nebraska is all Bush. Fixing my original post.

[Ats]

Me CD2: 90%
NH: 95%
NM: 35%
IA: 40%
MN: 60%
WI: 40%
OH: 50%
PA: 75%
FL: 25%

[Inverted Things]

Ats, your values give Kerry a 25.5% chance, with a 4.9% chance for a tie

[7,052,770]

Me CD2: 90%
NH: 95%
NM: 35%
IA: 50%
MN: 65%
WI: 35%
OH: 50%
PA: 85%
FL: 45%

[Inverted Things]

Me CD2: 90%
NH: 95%
NM: 35%
IA: 50%
MN: 65%
WI: 35%
OH: 50%
PA: 85%
FL: 45%

Chance for Kerry: 40.9%
Chance for Tie: 6.5%

[ilikeverin]

Me CD2: 90%
NH: 90%
NM: 40%
IA: 40%
MN: 60%
WI: 45%
OH: 55%
PA: 75%
FL: 35%

[Inverted Things]

Me CD2: 90%
NH: 90%
NM: 40%
IA: 40%
MN: 60%
WI: 45%
OH: 55%
PA: 75%
FL: 35%

Chance for Kerry: 34.1%
Chance for Tie: 4.5%