I've talked about this a number of times, but since Daniel Larison is asking (not necessarily to me, but in response to my previous response to him), I'll run through what I think again. I'll add at the start: what I'm talking about here are rough rule-of-thumb descriptions based on history -- postwar history, and especially postreform history (that is, 1972-2008).
To start with, I'd normally limit it to politicians who currently or recently served as either governors or Senators, or were recently on a national ticket. Are there exceptions? Sure. Well-known generals; maybe cabinet secretaries; maybe prominent Members of the House. Those aren't great platforms from which to run...it's probably a case-to-case kind of thing. I'd also add those who have run before and done reasonably well...Steve Forbes is more plausible in 2000 than 1996 How recently to do have to have served? Ronald Reagan won six years after leaving office; Richard Nixon, eight years out. That seems like about the outer limit...I can't think of anyone who has been out longer who really came close. No Captain Bateson class candidates, please. Again, there's some judgement here: someone who is ten years out but has been a prominent politician even out of office might qualify, while in my view Rick Santorum, who got clobbered for reelection in 2006 and was out of the spotlight since, doesn't.
OK, that narrows it down. Of the current Republican candidates (in or sort of in), it knocks out Bachmann, Cain, Roemer, Santorum, and Roy Moore, as well as apparent non-candidates Donald Trump and John Bolton. They simply don't have the observed credentials for winning a nomination.
The second big thing is that if you hold issue positions that will earn the opposition of important organized groups within the party, groups important enough to effectively hold veto power, then you're not going to be the nominee; you're also not going to become their presidential candidate if you oppose the bulk of the party, organized or not, on high-salience issues. This is even more of a judgement call than is the credentials part of it, and I'd suggest applying it lightly. Still, some cases are pretty obvious. The Republican Party as it currently has been for quite some time is not going to nominate a candidate who currently supports abortion rights, the Democratic Party will not nominate a pro-life candidate. In my view, this knocks out the libertarian candidates, Johnson and Paul; it also means that Jon Hunstman, with his moderate positions on several issues, and (perhaps more importantly) his service in Barack Obama's administration, is just not a plausible nominee.
It doesn't knock out candidates who would otherwise have been plausible nominees who have adjusted their positions in order to run. Mitt Romney may well have problems because of his past position on abortion, but Republicans nominated George H.W. Bush despite his flip-flop on the issue. I'm trying to distinguish here between candidates who are plausible and implausible nominees; of course, those in both pots are going to have additional strengths and weaknesses.
Now, one has to be careful with these things. At one time, it was probably true that only white Protestant men who had never been divorced were plausible nominees; none of those restrictions, of course, hold today. Issues, too, change over time -- and nomination battles are one of the main ways that parties change their positions, so surely we don't want to pretend that can't happen. We're only talking about what things appear to be now. But caveats included, still, it's not that hard to sort between those candidates who appear similar to those who won or at least came close to winning previous contests, and those who cannot claim any comps that came anywhere close to the big prize.
Now, a bit of clarification, in response to Larison's post. He writes:
Why are Pawlenty and Huntsman frequently referred to and treated as “main contenders” when Huntsman clearly isn’t and Pawlenty hasn’t shown that he is? The answer to that is that the rest of the field has been pre-judged as unacceptable to party elites for reasons of ideology or electability, but that’s not a very satisfying answer. Economic conservative activists declared Huckabee persona non grata early on, but it didn’t keep him from winning several contests and being competitive in a few others. McCain received almost nothing but “anti-endorsements” from many party and movement leaders, but these didn’t have that much of an effect. As I was just saying, the party elites’ declarations don’t carry nearly as much weight with primary voters as many seem to believe, and often enough the very things that make the rest of the field unacceptable to party and movement leaders are the things that make them popular with rank-and-file voters.There's very little I agree with here. Yes, (some) economic conservatives opposed Huck, but he had strong support from social conservatives. The results -- win in Iowa, competitive elsewhere, wound up falling well short of the ticket -- seem consistent with that. McCain? He wasn't popular early in the process, but wound up being acceptable to more or less everyone. I don't have data readily available, but I think this is selective memory; plenty of elites supported McCain. I think that both the endorsement and fundraising data would support that. The most unacceptable candidate to key Republican groups was Rudy Giuliani, and he of course went nowhere.
Remember, no one is saying that rank-and-file voters simply wait for endorsements and then automatically follow them. It's a more complex process. Endorsements and other positive buzz from insiders lead to candidates being taken seriously, which gives them more publicity, which helps them raise more money. As a results, candidates popular within the party get additional free and paid media, which raises their name recognition, which raises their horse race numbers...which then makes everyone even more likely to take them seriously, and it all starts again. Anyone can have a good media cycle, but background party support -- and an absence of strong opposition within the party -- helps that good cycle become far more productive.
Last bit...yes, it's still possible that a candidate with party support will turn out to be a dud with the electorate; I'd put Phil Gramm and, actually, Howard Dean in that category. Perhaps Tim Pawlenty will wind up in that category; it's far too early, however, to even begin to speculate about it. Check back again in six months. But winning without support from at least some party components? Not remotely possible.