Which Republican Senate seats will a Democrat win?

[Nym90]

My current prediction is PA, MT, MO, OH, RI, and TN, although the last two are probably a bit on the optimistic side.

[Nym90]

Anyway, I said I'd get to this later and I will.

Montana and Pennsylvania stand the highest chance of turning.  I give about an 80% of PA turning right now and about a 60% chance of Montana turning.  A lot about Montana depends on the nominee.

Missouri is the hard one for me to figure.  Talent's approval ratings are fairly strong and there is simply going to be very little "name recognition" bounce for McCaskill because she is already well-known.  But the closeness of the race at this point and the low Talent numbers make me uneasy in saying much.  I'm putting it at 35% chance of flipping, but this one probably has the most room to change.  It will either become ultra-competitive or relatively safe.

In Rhode Island, I think Chafee's biggest concern is the primary.  I think he's ahead right now and stands a 60%-65% chance of winning it.  If he wins the primary, I will probably put this race down in the Ohio category.  Right now, a 35% chance of flipping, slightly below Missouri

Which brings me to Ohio.  Do you believe Rasmussen or Mason-Dixon/uni poll?  Rasmussen, like SUSA, is going to be more turnout-based, since they do somewhat similar styles.  His one-day polling sample worries me though, and right now I would give much more cred to MD over the others.  I place Ohio at a 30% chance of going right now.

One of my general comments about the polling season so far is most of the polling has come from Rasmussen, which makes me wonder if we're not judging races too much on his polls.  Turnout-based samples tend to become more accurate towards election-time, but can be way off in the beginning, so I have a little bit of leeriness with giving too much credit to him.

I think Arizona has more potential to move than Tennessee right now.  Still, both are on the cusp of Lean/Likely on my lists and rate at a 15%-20% chance of movement (roughly the same as Maryland on the Dem list, for those keeping score).

Virginia is going to be tough.  You go with Webb, you could possibly snag conservative Indys and Dems, but DailyKos notwithstanding, I worry about the potential impact in NoVA.  Miller will obviously lose.  I think Allen could be kept under 60%, but under 55% will need a lot of luck right now.

Nevada is on the cusp of going safe for me right now.  Research 2000, a usually Dem-leaning pollster has it at 52%-32%.  Ensign probably gets the same amount as Reid did in 2004, maybe a tad more or less, depending on circumstances.

Like I said 3-4 is the most likely number here for now.  Above 5 is probably >5% at this time.

Sam, I'm not following...if I read you correctly you're saying that Democrats only have 40%+ chances in 2 states and yet are likely to take 3-4?

Otherwise, I'm with you on the analysis, though I would rate RI and MO closer to 50%.

I count 50% as being more likely than not chance of flipping.  If there are three seats with a 35% chance of flipping, it means to me that there's about a 15% chance of neither of the three flipping or all three flipping.  Though this is not mathmatically correct, it works in my mind. 

For example, on the Democratic seats, I would place Minnesota at the 30% level and New Jersey at the 25% level, so this translates to me that there's a 50-50 chance of flipping.

Hope I've confused you further. 

So in other words 35 percent actually means what most people would call 50 percent. Your mathematical system is very, um, interesting.

If there were truly a 35 percent chance of each of three seats changing hands, then there would be a 73 percent chance that at least one of them would change, and a 4 percent chance of all three changing.

If three seats were all at 50 percent, the chance of all three going the same way in either direction would be 12.5 percent, so at least your math isn't too far off in that case, once 35 is translated to 50.