Newt's surge peaked two weeks ago, with a 15 point lead over Romney. Last week, it was about a ten point lead. This week, it has been pretty stable, but is down to 4-5 points. The remarkably stable aspect of the data is Romney. He was at 22 when Newt peaked at 37. When Newt was 34 he got to 25. Since that blip, he's been between 22-24, and he sits at 24 on a rolling five day average and has for four days.
From that, the better question may not be whether Gingrich will plummet, but whether Romney is capable of rising above the mid to low 20s.
The numbers do have to eventually equal 100. At Newt's peak, he and Romney added to 59 (37-22). Then they did again at 34-25. Today, it's 28-24. The undecided has risen from 13 to 19 in that period, as I wrote almost a week ago. And they're staying undecided.
So if you think Newt won't win, you must think Romney can get into the thirties. If you allocate the undecided 19 percent among Romney and Newt, they both move into the thirties. But since the persistently undecided appear to be Cain or Perry voters who have resisted Romney, who predominantly identify with the Tea Party, and who ran from Cain to Newt and back to undecided, it's unclear why one would think not only that Romney could get some of these voters, but more of them than Newt will get.
The question remains: can 23% Romney get into the thirties? If he can't, the AntiRomney shall rise and rule the field.