Unfortunately, I haven't yet had the chance to read Michael Bailey and Forrest Maltzman's new book, but given that Maltzman's previous is probably the strongest example of a new wave of political science modeling of Supreme Court decision-making, I'm certainly looking forward to it. And I'm in general agreement with the bottom line of their argument at the Prospect: That the Court is less likely to rule that the Affordable Care Act is unconstitutional than the attitudinal model, which predicts Supreme Court outcomes based on general political affiliation, might indicate.
I'm a little less optimistic, though, with respect to the argument about precedent and how it might constrain the Court in this case. Without having read the new book (and hence seen their model laid out in detail), I'm obviously just going by what is presented in the piece. But while quantitative analysis of the Court is powerful and very useful, trying to determine the effects of precedent in this case is an example of some of its limitations. To borrow a point from this Mark Graber article—very much worth reading as an introduction to political science approaches to the Court—it is essentially impossible to disentangle "legal" and "policy" factors.
As Bailey and Maltzman themselves note, a major problem with trying to determine whether precedent will constrain Anthony Kennedy is that what precedents control this case is in dispute, and the mandate in the ACA is different enough from any prior set of facts that the Court could easily justify overturning it without overruling any previous case. Indeed, the "activity/inactivity" distinction was cooked up by litigators precisely to avoid conflict with previous precedents, and while I don't find this ad hoc argument remotely convincing—if I was the median vote on the Supreme Court the constitutionality of the ACA wouldn't be an issue. Kennedy and Scalia are likely to be much more receptive.
In addition, it's not clear to what extent Kennedy's greater willingness to uphold precedents can be distinguished from the fact that he's more ideologically moderate than his conservative brethren, and hence more comfortable with Warren and Burger Court precedents that Clarence Thomas despises. It's worth remembering that Kennedy wrote that Bowers v. Hardwick, which upheld a Georgia ban on sodomy, "was not correct when it was decided, and it is not correct today."
Bailey and Maltzman cite some important examples of conservative justices acting in unpredictable ways, but the other concern I have is that they generally involve conflicts between political goals and policy goals. In voting with the majority in Texas v. Johnson, which upheld the constitutionality of desecrating the American flag, Scalia may have been acting against the former, but he was acting with the latter (he's a fairly consistent First Amendment libertarian, and right now First Amendment libertarianism is as or more likely to help conservative litigants as anyone). Same thing with Gonzales v. Raich, which upheld a ban on homegrown marijuana: Kennedy and Scalia were acting against their legal policy goals but consistent with their political goals. Here, however, there's no conflict—the conservative political and legal policy goals are the same. This isn't really a case where we'd expect the conservative on the court to vote unpredictably. With respect to Rehnquist in Dickerson, since he didn't have the votes to overrule Miranda anyway it just suggests (as Bailey and Maltzman argue) that Roberts might be a sixth vote to uphold the ACA but won't be the fifth.
None of this is to say that the Court will strike down the ACA. Kennedy is a moderate enough that I can't predict his vote with any confidence. But I don't think that he will be constrained by precedent if he wants to strike the bill.