Cook vs. Sabato: why two top forecasters disagree on six Senate races

By Swati Singh · October 22, 2026 · Analysis

Dated editorial analysis; figures in this article are not recalculated automatically. Visit the linked race pages for current data.

Ratings as shown on Midterm Results Live on October 16, 2026

Two of the most widely cited election forecasters in the country look at the same 35 Senate races this year and mostly agree. But on six races, they do not. The Cook Political Report rates Iowa, Maine, Michigan, New Hampshire, Ohio and Texas as toss-ups. Sabato's Crystal Ball, published by the University of Virginia Center for Politics, rates all six as Lean Democratic.

That is not a small gap. Six Lean Democratic seats are the difference between a Senate that is "up for grabs" and one where Democrats are slight favorites. This piece explains what the rating labels mean, why the two outfits can reach different answers, and how the available polling lines up with each view.

What the labels actually mean

Both forecasters use a similar scale:

• Solid / Safe: The race is not expected to be competitive.
• Likely: One side is clearly favored, but an upset is possible.
• Lean: One side has an edge, but the race is competitive.
• Toss-up: Neither side has a clear advantage.

These are judgments, not probabilities. Neither Cook nor Sabato publishes a percentage chance for each race. A Lean rating does not mean "70% likely" or any other specific number. It means the forecaster thinks one candidate is ahead and would be mildly surprised by the other result.

That is important when comparing the two. A race that Cook calls a toss-up and Sabato calls Lean Democratic may reflect a small difference in judgment, not a big disagreement about the facts.

Where they agree

On most races, the two are in step:

• Alaska is a toss-up for both.
• North Carolina is Lean Democratic for both.
• Georgia and Minnesota are Likely Democratic for both.
• Kansas is Lean Republican for both.
• Nebraska and South Carolina are Likely Republican for both.
• Most of the remaining 20 or so seats are Solid or Safe for the same party in both.

The only other mismatch is the Florida special election, which Cook rates Solid Republican and Sabato rates Likely Republican. That is a smaller gap, but it points in the same direction as the other six: Sabato sees slightly more Democratic upside this year.

The six disagreements

Look at the polling column. In five of the six races, the most recent verified poll has the Democrat ahead. In Iowa, the Republican leads by two. So the polls lean a little toward Sabato's view in most of these races, and toward Cook's caution in Iowa.

But the leads are small. Only New Hampshire shows a gap larger than four points. In a single state poll, a lead of two to four points is well within the normal range of error. Cook's toss-up rating is a reasonable reading of numbers that close.

Why forecasters can see the same race differently

1. How much weight to give the national mood

National indicators this fall are poor for the party in the White House. In the polls in our ledger, presidential approval ranges from about 30% to 42%, and the generic congressional ballot ranges from a two-point to a 16-point Democratic lead. A forecaster who puts more weight on that national environment will move close races toward the out-party. A forecaster who puts more weight on each state's history will be slower to move a Republican-leaning state like Texas or Iowa.

2. How much weight to give state history

Texas has not elected a Democrat to the Senate since 1988. Iowa has voted Republican in recent statewide races. Ohio has moved to the right over the last decade. Calling any of those Lean Democratic is a statement that this year is different. Cook's toss-up label effectively says "maybe, but we need more proof."

3. Candidate quality

Several of these races feature candidates with unusual profiles. In Texas, the Republican nominee is Ken Paxton, who defeated incumbent John Cornyn in the primary. In Ohio, the Democrat is former senator Sherrod Brown, who is well known to voters statewide. In Maine, Susan Collins has won re-election before in years that were bad for her party. Forecasters can reasonably disagree on how much each of these factors is worth.

4. Timing and style

Ratings are updated on different schedules, so one outlet may simply have moved first. The Crystal Ball has also historically issued final ratings before election day that place every race on one side or the other, with no toss-ups left. That habit means Sabato tends to commit to a lean earlier and more often than Cook.

Which one should you trust?

Neither one is "right" until the votes are counted. Both have long track records, and both have missed races in past cycles. The most useful way to use them is together:

• When Cook and Sabato agree, the rating is a strong signal.
• When they disagree, the race is genuinely uncertain, and polls and early results deserve extra attention.

For the six races above, the honest summary is this: Democrats have a small edge in most of the available polling, but not a big enough one for every forecaster to commit. That is what a close Senate map looks like two weeks before an election.

You can compare both sets of ratings on our Senate page and our ratings map, and see how we choose which polls to include on the methodology page.

Relevant race pages: Alaska Senate · Florida Senate · Georgia Senate · Iowa Senate · Kansas Senate · Maine Senate · Michigan Senate · Minnesota Senate · Nebraska Senate · New Hampshire Senate