If you’re trying to predict how many U.S. Senate seats the Democrats will win in November, you definitely want to look at all the latest polls. You also want to weight the credibility of the polling outfits, and FiveThirtyEight is helpful in this because they provide grades and explain historical partisan skews. When assessing individual races, you’ll also want to make a guess about whether Biden or Trump will carry the state in question, and it pays to keep an eye on how well candidates are funded and what kind of outside help they’re getting. Finally, you want to know the challengers and make an assessment of their fit for their state and they’re basic political skills.

Looking at all the Senate polls released since June 1, I see that the Democrat is favored in one or more survey of Arizona, Colorado, Montana, Kansas, Iowa, North Carolina, both Georgia races, Kentucky, and Maine. That’s ten races. All of these should be considered potential pickups, although I remain skeptical about Kentucky. Against this, Senator Doug Jones of Alabama has trailed both Jeff Sessions and Tommy Tuberville. Getting a net gain of eight seats therefore seems quite possible for the Democrats. There have also been surveys indicating that South Carolina will be a nail-biter, and both Texas and Alaska look like they’re potentially competitive. So far, however, no poll has shown a Democrat leading in those three states.

Most prognosticators are very conservative about predicting that incumbents will lose, and since “the presidential result and the Senate outcome matched up in every state four years ago,” they’re also extremely reluctant to go out on a limb and suggest that a Democratic Senate candidate will win in any state that Trump carries. This is why they’re not saying that the Iowa and Montana races currently lean Democratic despite several recent polls showing the Democrats leading there. They’re just not confident that Trump will lose either of those states or that there will be enough ticket-splitting to provide a mixed outcome.

To me, this is both lazy and cowardly. It’s lazy because there are some states, like Montana or Alaska, where a mixed outcome is not particularly unlikely, and other states like Alabama and Georgia where that result is hard to envision. A lot depends on how elastic the electorate is, and it’s least elastic in the Deep South. Montana and Alaska, however, have split their tickets in recent memory, and Iowa sent both Tom Harkin and Chuck Grassley to the Senate for decades. You shouldn’t put too much emphasis on the perfect Senate/presidential alignment of 2016.

It’s cowardly because people don’t want to know the current state of play. They can read polls just as well as the experts. They look for analysts to give them something more, and if an analyst won’t say a race leans one way until the polls confirm it, then they’re not providing any analysis at all.

For one thing, Trump is on a clear downward trajectory, and the Senate polls reflect that. Do we have solid reasons for believing that this trajectory will be arrested or even reverse?

At Roll Call, Nathan Gonzales says no: “It seems unlikely that there would be a news event that would dramatically improve the president’s standing and thus boost GOP prospects.”

So, why then, does he persist in calling Maine a “Toss-Up” when Biden leads there in the most recent poll by eleven points and Susan Collins has trailed in every head-to-head survey released this year, save one?

Likewise, Biden has led in 12 out of 14 North Carolina polls since early June, and Cal Cunningham has been ahead of Thom Tillis in eight out of ten and tied in another. Yet, Gonzales still has the Tillis race rated as a “Toss-Up.”

This isn’t to suggest that these races are already decided. But we’re talking about making predictions. In that sense, rating something too close to call is already a punt, and it’s only justifiable if you have some solid reason to believe that the race will be heading for recount territory. Based on the current trajectory, which seems unlikely to reverse, any state where both Biden and the Democratic Senate candidate are currently leading should be rated lean-Democratic, and the same can be said for any states that are even or where the Republicans are only narrowly ahead.

To see what I mean, I am not predicting that Senate Republican leaders Mitch McConnell and John Cornyn will lose, but for two completely different reasons. I expect Cornyn to run substantially ahead of Trump in Texas and McConnell to run substantially behind Trump in Kentucky. Texas is too close to call at the presidential level although I think Biden can carry it in a fair election. Even if he does, though, I think Cornyn should be favored to fare better as he’ll do a better job of holding onto traditional conservatives in the suburbs. If Cornyn goes down, it will surely be Trump and Governor Abbott’s bungling of the COVID-19 crisis that sinks him, but things will have to be really bad. In Kentucky, however, Trump is polling 17-20 points ahead of Biden. Mitch McConnell is very unpopular in his home state, but I don’t think there are enough Trump/McGrath voters in the Bluegrass State to cost him his job. McConnell should begin to worry when he sees Trump with single-digit lead, but not before then. My guess is that will never happen in Kentucky, but I reserve the right to change my mind.

For me, the hardest races to predict are in South Carolina and Georgia. My guess is that more people in Georgia will have the intent to vote for Biden and the Democratic Senate candidates, but that’s not good enough in a state that makes an art form out of cheating. The South Carolina race is just going to be close, and it’s going to take quite a bit more deterioration in Trump’s fortunes before I’m convinced that Lindsey Graham is the underdog there.

As of today, I’m prepared to predict that the Democrats will ultimately carry the Senate races in Montana, Kansas and Alaska. They’ll also win the races in Arizona, Colorado, North Carolina, and Maine. They’ll win both races in Georgia, but may not be awarded the wins, and Doug Jones will lose his seat in Alabama. That’s a net gain of eight seats that might only count for six. So, we’re looking at a Democratic majority of either 55-45 or 53-47, depending on the fairness of the Georgia results. The next to fall in order of likelihood are South Carolina, Texas, and Kentucky. If the pandemic gets really out of control or Trump takes his screw-ups and criminality to the furthest reaches, the Mississippi race could get interesting. I don’t think Tennessee or any of the other races have much potential in any scenario, but maybe I’m selling my imagination short.

In a true cataclysm, the Democrats might win as many as 59 seats, but I think the 53-55 range is the likeliest outcome.