The model uses situational variables such as yards per pass attempt on first down, percentage of first downs that a team passes on and yards per rush allowed on third down. The situational data allows us to examine specific match-ups and understand why one team has a higher probability of winning than another. In order to compare teams fairly however, we have to compare their performances against a "standard" team. While no team is perfectly average in every category, I constructed the average team to have a constant barometer to compare teams. With that very brief introduction, the rankings for the upcoming season are below. I included a ranking for each team, their estimated points against an average team, and their rank from the end of last year according to the NFL (who use yards per game to rank offenses).
The biggest change from the end of last season to the beginning of this season is for the team at the very top of the rankings: San Diego. They jump from the 10th best offense last season to the best offense this season. There is a caveat to this jump in offensive ability however, and it comes from Norv learning from his past mistakes. If they just pass a little more frequently on first down, using one of their best assets, they will be able to tear up most defenses.
Condolences to fans in Cleveland however, as, even with some assumed regression to the mean, the Browns are not likely to get out of the cellar. They are projected to repeat as the worst offense in the league. This low ranking is driven by the incredibly inept passing game that was more than 1 yard below league average per attempt on first and second downs. With 80% of their roster returning, it does not look like there will be much room for improvement.