Republican presidential candidates participate in the first debate of the GOP primary season in Milwaukee. TNS

It was an unprecedented week in US presidential nomination politics, but a lot of pundits remain certain that all roads still lead to Donald Trump being the Republican candidate. As analyst Charlie Cook put it: “Barring Trump having some health-related problem before the convention, this nomination is settled.”  But polling following the first debate among Republican presidential hopefuls shows that while Trump is the most likely nominee, there’s still plenty of uncertainty.  Let’s start with the debate polling from FiveThirtyEight/Washington Post/IPSOS. Three candidates stood out as the winners according to those who watched the event, with 29% selecting Florida Governor Ron DeSantis as having the best performance, followed by Vivek Ramaswamy at 26%, and former UN ambassador Nikki Haley at 15%.

The key finding in the survey, however, was that each of the eight candidates gained in a critical question: whether debate-watchers would “consider” voting for them in upcoming primaries and caucuses. It’s a good question because the debate to some extent simulates the effects of a campaign. Remember, most voters have barely paid attention to nomination politics so far. The first event in Iowa is still months away, and most Americans won’t be voting — or exposed to heavy advertising and campaigning until February and March. With the exception of Trump and perhaps former Vice President Mike Pence, these are not yet familiar candidates to most voters.

Some of the increases were marginal— former Arkansas Governor Asa Hutchinson’s improvement to 9.4% from 8.5% on that question hardly makes him a threat to anyone. But immediately after the debate, 50% of those polled said they would consider supporting Haley, compared to 30% who said they would before the event. And those who said they would support DeSantis increased from 63% to just over two-thirds, a smaller but perhaps even more consequential increase. Meanwhile, all three candidates the pollsters asked about who weren’t on stage saw their support drop, including Trump, who lost about five percentage points and fell below DeSantis on that measure.

That’s not all. As political scientist Meredith Conroy pointed out Thursday, “Candidates willing to go after Trump last night (Haley, Christie, Pence) didn’t get punished.” Indeed, not only did all three expand the group willing to consider voting for them, but their net favorable ratings all went up. Christie and Pence still remain underwater by that measure, and Haley’s criticism was limited to Trump’s electability, still if the candidates conclude that they can gain support from attacking the former president, it may increase their willingness to do so. It’s true that no candidate has held as large a polling lead as Trump has and failed to win the nomination. But that’s a less impressive statistic than it seems.

For one thing, there aren’t that many cases of big early leads for open nominations — excluding incumbents running for reelection— during the 50 years of the open nomination system.  We can also be sure that at least some of Trump’s large lead is air that he’ll lose once voters get to know other candidates. There’s plenty of history of the best-known candidate doing especially well in early polling and then losing some or all of that lead once voters start paying attention in the days and weeks before they actually go to the polls.

The key thing to know about voters and nomination politics, however, is what Nate Silver said. “Most primary voters like multiple candidates, and that makes multi-candidate primaries intrinsically volatile.”  He’s correct. Pundits who marvel at how many Republican voters still like Trump despite the long list of reasons not to are getting it backwards. We should expect most of a party’s voters to like most of the party’s politicians. What’s important about Trump is that many Republicans don’t like him, and that there’s even some evidence that being indicted has eaten away at his support within the party. But Republicans don’t actually have to dislike Trump for him to lose the nomination. They just have to like another candidate better, and evidence from the debate suggests that once campaigning begins in earnest they’ll find other candidates that they like.

And don’t worry too much about potential winnowing failures — the possibility that several candidates will split the anti-Trump vote. That effect was real but overrated in 2016, and we’re already seeing signs that winnowing may work normally in this cycle. Republicans have already effectively narrowed the candidate field to eight Trump opponents by excluding others from the first debate, and may squeeze one or two more out after raising the threshold for qualifying for the September debate. More will likely follow and if they don’t formally drop out it’s hard to see how candidates polling at 1% or less will matter one way or another.

To be sure, Trump is absolutely the most likely to win the nomination. But it’s simply too early to be sure and one can’t repeat too many times that everything about this is unprecedented.

It’s hard to know how to extrapolate the numbers. Presumably some Republicans who are certain to vote for Trump wouldn’t bother watching primary debates. But voters who don’t pay much attention to politics, a group that tends to have less solid preferences, are also less likely to watch an August debate.  The exception here is Hutchinson, who moved from being unpopular but barely known before the debate to better known (among debate watchers) and even more unpopular afterwards. The likely lesson here is that candidates known mainly for attacking Trump - and Hutchinson had little time to do much more Wednesday night - are unlikely to be popular among Republicans, but a more well-rounded candidate such as Haley or perhaps even Pence might be viable.  It’s certainly possible to imagine a situation where two or three candidates stay in and collect enough votes that Trump is nominated despite failing to reach 50% of the vote, just as it’s always possible to come up with a scenario that produces a contested convention. But the logic of winnowing is a strong one, and anti-Trump Republicans are especially attuned to it in this cycle.

Jonathan Bernstei, Tribune News service