From Public Policy Polling’s most recent poll of 814 likely voters over 10/30-10/31 (PDF), which shows respectable leads for Mark Kirk and Bill Brady:

No matter who wins either of these races Illinois voters will be left with a Governor and Senator that they don’t like. Giannoulias’ favorability is 35/49, Kirk’s isn’t much better at 39/45. Quinn’s approval is 32/54 and that’s a good thing for Brady because voters don’t like him either, giving him a favorability of 39/45.

Looking at the crosstabs, this is sort of interesting:

Who Brady voters voted for in 2008: 84% (McCain) 16% (Obama)
Who Quinn voters voted for in 2008: 9% (McCain) 67% (Obama)

Who Kirk voters voted for in 2008: 85% (McCain) 17% (Obama)
Who Giannoulias voters voted for in 2008: 6% (McCain) 72% (Obama)

In short: everything sucks but more so for Democrats, so Giannoulias and Quinn are almost totally reliant on party-line votes.

Here’s what I’m thinking right now. The numbers are pretty comparable between the two races. Giannoulias isn’t a very good candidate—though I think Giannoulias made a sensible decision to run as a relatively liberal Democrat, whatever that’s worth—and Kirk is much more appealing to moderates and conservative-leaning Democrats than his gubernatorial counterpart. So I think Kirk wins on that account.

Meanwhile, Quinn is a more appealing candidate than Giannoulias—though his brief tenure as governor in the wake of Blagojevich’s incompetence, hamstrung by a feckless Democratic legislature he can’t run against, hurts him—and Brady is less likely to pull in fence-sitters than Kirk. But Scott Lee Cohen (money) and Rich Whitney (limited name recognition) combined will make the difference for Brady.