Unless between now and December 12th something ugly happens for Roy Moore, like solid proof that Moore is a sexual predator, he is going to win the senate race and I do not think it is going to be a close race.
Mitch McConnell suspects this:
The RNC know this: http://www.cnn.com/2017/12/04/politics/rnc-roy-moore-alabama/index.html
And most importantly, you can see it in the polls. Most of the latest polls listed on realclearpolitics show Moore is winning, except for a Washington Post poll from the end of November and a Gravis poll that was conducted from December 1st to the 3rd The RCP average is Moore +2.3, but I imagine he should be at least 5 points ahead.
The Gravis poll is not a proper indicator of how the race will go, let me explain why.
First we have to understand what the electorate in Alabama looks like. Since the turn of the century, Alabama has not voted for a Democrat governor, senator, or president. Don Siegelman was elected governor in 1999, he was the last. The following table and chart shows some voting break downs for recent statewide elections:
*Note: Sessions (1996) and Shelby (1998) are included to show their first win to the Senate.
The above table and chart show the Republican party is getting stronger and stronger in Alabama. Trump received more votes than any previous Republican Presidential candidate and won by a wider margin than any of his recent predecessors.
Furthermore, Trumps approval rating in September (in Alabama) was at 59, this rating makes some sense, since it aligns closely to the percentage of the vote he received:
Another thing to keep in mind, Alabama is so Republican that in 2014 the Democrats did not even run a candidate for the senate: click here to see
Equipping ourselves with this information, lets dig into the aforementioned poll and find out if it holds water.
The Gravis poll shows Roy Moore is behind by 4 points. This poll is light on details, but it does have a “do you favor Trump” rating and an approval rating for President Trump. Both of those ratings are much lower (51% approve, 52% favorable) than the 2016 voting results or the above mentioned approval rating for Trump in Alabama. That seems odd.
This tells me that the sample is heavy on pro-Jones voters. If you were to adjust it to reflect the actual Alabama electorate, it would swing to show Roy Moore up by 5 to 7 points, similar the most recent poll showing Moore up by 7: