Well, given the stunning success of
this model
during the 2004 Presidential Election, we'll see if it can withstand
the rigors of this year's mud-wrestling. Turn OFF the TV
&
radio; just check the web-site!
For details of the model, please see its description on the
2004
Model Page. The basic concept is the same, except
only the
Class
I Senatorial races are being considered in the
betting. That
means only the incumbents (17 Democrats, 15 Republicans and 1
Independent) form the basis for the daily calculation. In
other
words, the
Class
II and
Class
III Senators (terms expiring in 2009 and 2011, respectively)
form a
constant population of 27 Democrats and 40 Republicans (there are no
Independents in either of these classes), with the remaining 33 seats
in contention.
Without further ado, here is my
daily
predicted outcome .
Happy
Voting!
November 8th, 2006 recapitulation--
The 110th Congress Senatorial final results are in:
So it looks pretty interesting. Beginning 30 days in advance of the election, the wagering-based representation of voter
intent
showed a steady split of 49:49:2 (Republican : Democrat : Independent)
for six days. About 21 days before the election, modest
excursions skewed slightly to 51:47:2 for about 13 days. The
ratios then returned to 49:49:2 for the final 5 days up to and through
the election. On the day of the election, low amplitude
oscillations occurred (probably driven by frantic contract trading)
before settling to the final 49:49:2 result, a curious coda!
See this
site for a proper summary of the 2006 Senate race results.