I don't like to bet (esp. with money), except for a rare lottery ticket or two and a rarer slot machine playing for fun.
But, I happened to check the online betting site Intrade for the 2008 election and found something very interesting. Link (click the "2008 US Election" link to the left).
They have the market prices for wins in the primary and the Presidency. For example, Hillary Clinton is trading at (bid):
2008DEM.NOM.CLINTON: 71.6
2008.PRES.CLINTON(H) 47.6
Therefore, roughly speaking, the participants of Intrade odds market think that:
-- Hillary Clinton has a 71.6% chance of winning the nomination.
-- and she has a 47.6% chance of winning the Presidency, all told.
So, let's ask the question: what are HRC's chances of winning the general election if one assumes that she wins the nomination?
This quantity is a case of the notion of conditional probability. From the wiki:
Conditional probability is the probability (or chance) of some event A (taking place), given the occurrence of some other event B. Conditional probability is written P(A|B), and is read "the probability of A, given B".
In our case, if A is the event of candidate X winning the general, and B is event of X winning the primary, then P (A|B) is the probability/chance of X winning the general given a primary victory, which can be considered to be a measure of "electability," arguably.
There is a related formal quantity called the joint probability:
Joint probability is the probability of two events in conjunction. That is, it is the probability of both events taking place. The joint probability of A and B is written P(A,B).
Since for a major party candidate, to win the Presidency, he/she needs to win both the primary and the general (3rd parties excepted), in our example, P(A,B) represents the probability of X winning the Presidency, i.e. that of winning the nomination and then winning the general.
There is a simple formula that relates all of these quantities:
P(A|B) = P(A,B) / P(B)
Applied to our case, in other words: the chances of winning the general given the nomination = the ratio of winning the presidency over that of winning the primary.
Working out the number for HRC's case, we get:
the chances of HRC winning the general given the nomination =
the ratio of the chances of winning the presidency over that of winning the primary =
47.6%/71.6% = 66%
Let's give the quantity "chances of winning the general given the nomination" a name: the Electability Index (and "Intrade Electability Index" when applied to numbers in a given snapshot at Intrade).
No, I don't like the word electability much, because it's been much abused in both the previous and the current presidential cycles. But, at least we have a solid definition of the word now!
I compiled the tables for 6 candidates from both of the major parties (plus Gore) from the snapshot of the Intrade market that I've found, and computed the Intrade Electability Indexes for each them. Here is how the numbers shake out:
As one can see, the table reveals some very interesting numbers.
On our side, the market currently thinks that Gore has the highest chance of winning the general election if nominated: 98%! Please see this diary by NYPopulist for some recent GE actual poll numbers for Gore.
Hillary and Obama follow at 66% and 60%, respectively. Edwards comes in (maybe surprisingly low or high depending on the person you ask) at 56% (I suspect that this number dropped for Edwards as a result of his decision to accept public funds.)
On the GOP side, Ron Paul tops the list at 55%, followed by McCain on our "Electability Index" scale here. I'm a bit surprised to see Huckabee's electability index come at a low 19%, but as one can see at the link, he tops the list of VP prospects for the GOP at 21% odds. Obama (21%) and Bayh (20%) top the VP list from our side.