August 13, 2008
By ChimbaChumba
The Pollsters & ChimbaChumba’s Poll of PollAverages
Pollster 46.5 - 43.3 (+ 3.2%) (y’day +2.9%)
CNN 47.0 - 41.0 (+ 6.0%) (y’day +5.0%)
RCP 46.8 - 42.2 (+ 4.6%) (y’day +4.8%)
538 49.9 - 47.8 (+ 2.1%) (y’day +1.9%)
ChimbaChumba’s Poll of PollAverages: Obama +4.25% (yesterday +3.65%)
Betting Market (Betfair)
Obama $1.51, McCain $3.15 (y’day - Obama $1.53, McCain $3.20)
The Handicappers - Obama’s Win Percentage
Betfair ($1.51 adjusted to 100% price) - 67.44% (y’day 67.5%)
538 Electoral Predictions - 65.8% (y’day 63.5%)
Rasmussen Balance of Power Calculator - 60.5% (no change)
Ohio & Colorado watch (key swing States if race is to be close)
Obama’s lead Ohio Colorado
RCP Ohio +0.5 (no change) +1.6 (no change)
Pollster Ohio +3.6 (no change) +2.5 (no change)
538 (trend-adjusted) Ohio +1.8 (y’day +1.1) +2.4 (y’day +2.2)
ChimbaChumba’s Poll of PollAverages: +1.96 + 2.16
538 Projected Win % 59% - 41% ( y’day 54-46) 65% - 35% (y’day 63-37)
ChimbaChumba says - Obama gains with all the “pollaveragers” today, taking Obama’s lead in the ChimbaChumba pollaverage to an all-time high of 4.25%. Obama also records an uptick in our daily watch-States Ohio and Colorado. The betting market moved in sync, indicating that punters are following the polls.
I am a little sceptical of the validity, or relevance even if valid, of this improvement in Obama’s numbers. 538 calls today the “biggest polling day in several weeks” with 7 new polls out. The only polls from key or battleground States are polls from Nevada(1) and Virginia(2). In Nevada, Rasmussen’s latest poll has McCain going from 3 points down to 2 points up. In Virginia, McCain holds steady at +1 in the Rasmussen poll, and was +0.5 in a 2nd Nevada poll by InsiderAdvantage.
So on the one hand the net effect of these 7 polls, and whatever other stuff 538 feeds into its numbers (and I don’t mean that critically - 538 is the best info available imo), has caused 538 to increase Obama’s national lead in “popular vote” and “winning percentage”.
On the other hand, in 2 key States where there was new polling data specific to those States, 538 re-rated McCain to 62% favourite in Nevada and 54% in Virginia. Based on what I assume is general data, McCain slips to 41% underdog in Ohio and 35% in Colorado. Remember, if McCain were to win all 4 of these States, based on the current Rasmussen Electoral College State by State balance of power table, McCain would win the election
So it seems to me there is some discord here. Specific polling in 2 key states was a small fillip for McCain, yet the other indicators (polling and betting) have trended solidly in Obama’s direction.
What if there is a poll in Ohio next week which showed that State was an even race (the last specific Ohio poll from Quinnipiac only had Obama +2)? I think it would cause McCain’s price to tumble to about $2.70.
This highlights the absence of relevant data to form an opinion, National polls may be misleading in a close race. Of course the average punter is not helped by misinformation either - in Britain’s “No.1 quality newspaper”, The Telegraph, we can read the following from Irwin Steltzer today. Which polls is this idiot referring to?
The questions that hang over Barack Obama
By Irwin Stelzer
“But the polls suggest that the candidates are in a dead heat”.
Comments
