The latest chart from pollster.com estimates Obama's lead in PA at no less than 15.3% (53.8 to 38.5).
To give a sense of how wild it looks for McCain to be heavily targetting this state at this stage of the election, let's compare Red states with comparable or lower leads.
We all know that Obama is either targetting or believes he has already effectively won the following Bush 2004 states:
IA, NM, CO, VA, FL, NV, OH, NC, MO, IN.
But a look at the pollster.com site quickly shows that there are many more Bush 04 states which in terms of the polling gap look more winnable for Obama than PA appears to be for McCain.
Here they are, in descending order of winnability:
ND: O +3.6
MT: O -3.7
GA: O -5.4
MS: O -9.5
AR: O -10.6
WV: O -10.7
SC: O -11.5
LA: O -12.2
AZ: O -12.5
KY: O -14.1
Finally, the closest state to PA in terms of Obama's deficit is (drumroll...)
TX: O -15.6
So, in other words, the equivalent strategy to the one McCain appears to be adopting would be if Obama pulled out of states where he's moderately behind (like, say, GA) to concentrate most of his resources instead on winning TX.
Yes, indeed, those 34 EV would be very attractive and could indeed give him more paths to a (landslide) victory, but would it really be sensible to commit a lot of resources there when there are other apparently more winnable states on the list?
To be fair to McCain, his options are really limited at this point, but he's only 5.7% behind in Colorado which has a much redder lean in terms of party ID. He's about as likely to win Colorado as Obama is to win Georgia; whereas he's about as likely to win Pennsylvania as Obama is to win Texas. I'm no bigshot campaign strategist, but I know which option I'd go for.