The NY Times 538 Blog is by far the best source right now for in depth analysis of the latest polling and a detailed look at how it relates to delegate math.
In this article, he breaks down in-depth why a) Illinois is absolutely vital for Santorum to have a shot at the nomination and b) Why people like me are so certain the Romney is going to get the nod, one way or another.
Even if Santorum does substantially better in the polls than he’s doing now (5% better across the board) and manages to win Illinois, Delaware, Montana and New Mexico (where he’s currently behind), he’ll STILL be behind in the delegate count. In that scenario, Romney might not be able to get to 1144, but:
Does that mean we would be headed to a brokered convention? In my view, probably not quite. There would still be enough superdelegates and other unpledged delegates to put Mr. Romney over the top; he would need the support of about two-thirds of them.
My guess is that he would probably get it, mostly because there was no better alternative. Mr. Romney would need a lot of help in this case. But he would not need as much help as Mr. Santorum. Even if Mr. Santorum won all the unpledged delegates, plus all of Mr. Gingrich’s delegates, that would leave him at 1,127, still just shy of a majority. So a Santorum nomination would require a brokered convention, while one for Mr. Romney would not. That is a pretty powerful reason for the unpledged delegates to align behind Mr. Romney.
To me, this math would suggest that it’s actually in Santorum’s interest for Gingrich to stay in the race. If Gingrich gets out, even if he pledges his delegates to Santorum, some percentage (perhaps as many as 25%) of his supporters would go to Romney instead of Gingrich. The 538 author argues the opposite, but I would argue that Santorum’s best chance is at a brokered convention – and the best chance of a brokered convention is keeping the race as fractured as possible.