Analytics expert Cynthia Frelund took data from the past 10 NFL campaigns to create historical references for personnel, scheme and matchups, identifying factors that are proven to lead to wins (or losses). She vetted the correlations with a bunch of real football people (e.g., coaches) and had her math checked out by real math people (e.g., PhDs) to make sure the model reflected reality as much as possible. Then she compared this season's personnel, schemes and matchups -- with the vetted mathematical weightings -- and simulated the season to produce a ceiling, floor and projected win total for all 32 teams. The ceiling-to-floor range is something to home in on -- a big discrepancy means the team is projected to play in more close games.
Now, with the addition of a 17th regular-season game, this is the biggest NFL season ever. Inherently, Cynthia ran the most simulations ever: 300,000 runs of every single regular-season game, which equals 81,600,000 total games "played."
Without further ado, here are the projected win totals for AFC teams in the 2021 campaign, ordered from most to least wins, with playoff berths noted. Click here for NFC win totals.
NOTE: All FanDuel over/under listings are current as of 2 p.m. ET on Monday, Sept. 6.
Ceiling: 14.6
Floor: 9.9
FanDuel over/under: 12.5
After an offseason rebuild of the offensive line, Patrick Mahomes forecasts to pace the league in passing yards, earning at least 5,000 yards in 58.7 percent of regular-season simulations. In my models, generally speaking, 58 percent has historically been a great benchmark.
Ceiling: 13.3
Floor: 9.2
FanDuel over/under: 11
Teams with a lot of player continuity -- and in the Bills' case, coaching continuity, as they retained both offensive coordinator Brian Daboll and defensive coordinator Leslie Frazier -- forecast to start off the season strong. This also adds more certainty to my model's prolific projections for Josh Allen (more than 4,500 passing yards in 57.9 percent of season simulations) and Stefon Diggs (more than 1,350.5 receiving yards in 59.2 percent of simulations).
Ceiling: 12.6
Floor: 8.9
FanDuel over/under: 10.5
No offensive line has a higher win-share projection than Cleveland's, but that isn't the biggest reason why the Browns are projected to jump from third place in 2020 to AFC North champs this year. The real driving force behind this rise: an improved secondary. At the end of 2020, Cleveland's defensive backs ranked 25th in the NFL in win share. Entering this season, the unit projects to rank fourth. Also, Nick Chubb forecasts to score at least 12 rushing touchdowns in 55.9 percent of simulations.
Ceiling: 12.9
Floor: 8.5
FanDuel over/under: 11
Injury issues to start the season -- most prominently, J.K. Dobbins' ACL tear -- dropped Baltimore's median win total to 10.8 and added to the range (volatility) between the floor and the ceiling. Thus, Cleveland opens as my model's AFC North favorite. On the plus side: Don't sleep on TE Mark Andrews, especially in the end zone, as the fourth-year pro earns more than 7.5 touchdowns in 59.5 percent of simulations.
Ceiling: 13.1
Floor: 8.5
FanDuel over/under: 9
The addition of Julio Jones creates a higher likelihood of Derrick Henry facing fewer stacked boxes this season, making him the favorite to lead the league in rushing (again). "King Henry" earns at least 1,539.5 rushing yards in 59.9(!) percent of simulations. When Henry isn't contacted until he has reached his full speed, the 247-pounder earns an average of 2 yards more than the next-closest back.
Ceiling: 11.9
Floor: 7.7
FanDuel over/under: 9.5
There is a bit of a logjam in the win projections between the Dolphins (9.6), Pats (9.4) and Colts (9.3), who all rank between sixth and eighth in the AFC, suggesting one will miss the playoffs. Consequently, the Week 1 matchup between the Dolphins and Patriots looms large right off the bat. While the game is in Foxborough, my simulations have the Fins upsetting the Pats in 56.8 percent of results.
Ceiling: 11.8
Floor: 7.5
FanDuel over/under: 9.5
When the Patriots released Cam Newton and handed the starting reins to Mac Jones, New England's win total ... didn't really budge. This is not to say the rookie first-rounder has equal value to the former league MVP from Game 1; rather, the sum of the parts on this team with the clarity of Jones starting didn't drive a major shift in the win total. However, this did make earlier games in the season less-likely wins and later ones more-likely Ws.
Ceiling: 12.2
Floor: 5.5
FanDuel over/under: 9
The spread between the floor and the ceiling is massive here at nearly seven games. A big part of this volatility? The offensive line -- specifically, LG Quenton Nelson and LT Eric Fisher -- already being banged up. This leads to Jonathan Taylor not crossing the 1,200-yard mark in 58.9 percent of simulations.
Ceiling: 11.2
Floor: 6.9
FanDuel over/under: 9.5
How the Chargers start the season will determine if this team will contend for a wild-card spot. L.A. has an early bye in Week 7, but before that, the Bolts face Washington, Dallas, Kansas City, Las Vegas, Cleveland and Baltimore. Brandon Staley's defensive game plan -- specifically, the team's ability to consistently bring pressure -- is a major driver of game outcomes. This offense, with Joe Lombardi at the helm, features Austin Ekeler earning at least six rushing touchdowns in 56.0 percent of projected outcomes.
Ceiling: 11.3
Floor: 5.9
FanDuel over/under: 8.5
While the Broncos' projected win share puts them at 10th in the AFC, their potential to make the playoffs ranks seventh-best. Wait, what? Remember: Projections factor in a percentage chance each team wins each game (say, 60/40, for example), but then once a game is won, it turns to 100/0. The Broncos actually have more higher-probability single-game wins than some other potential AFC playoff teams. They also have the second-best odds of starting the season at 3-0 in the entire NFL, with an opening stretch of at Giants, at Jaguars and home vs. Jets.
Ceiling: 9.9
Floor: 5.0
FanDuel over/under: 8.5
In 58.5 percent of outcomes, Ben Roethlisberger logs at least 4,000 passing yards. In 56.9 percent of season simulations, rookie Najee Harris scores eight or more rushing touchdowns.
Ceiling: 7.9
Floor: 2.8
FanDuel over/under: 7
The inability to consistently generate defensive pressure on passing downs combined with the potential for low O-line production creates unfavorable conditions for high-probability success. However, the Raiders' big-play potential on offense (passes of 20-plus yards, rushes of 10-plus) is higher than last season, when they finished tied for seventh in the NFL with 114.
Ceiling: 6.8
Floor: 2.1
FanDuel over/under: 6.5
The Bengals have my model's 28th-ranked offensive line. Not ideal. On the plus side, rookie receiver Ja'Marr Chase forecasts to score at least seven receiving touchdowns in 57.0 percent of simulations, while Joe Burrow's median passing projection (4,403 yards) has him netting the most throwing yards in the AFC North.
Ceiling: 6.9
Floor: 2.5
FanDuel over/under: 6
Zach Wilson projects to lead all rookie quarterbacks in passing yards (3,907) and passing touchdowns (23).
Ceiling: 6.6
Floor: 2.0
FanDuel over/under: 6.5
No. 1 overall pick Trevor Lawrence forecasts to rush for at least four touchdowns in 57.2 percent of simulations. But as far as total touchdowns go, it's more likely that Jacksonville's run game as a whole will be a source of red-zone success, meaning Lawrence's total touchdown number falls at 28 or fewer in 57.0 percent of simulations. This is not at all a knock on the rookie, but rather a projected roadmap of how scores are most probable to occur, given this team's composition and Urban Meyer's past play-calling tendencies.
Ceiling: 5.4
Floor: 1.1
FanDuel over/under: 4
The extreme amount of uncertainty surrounding this team creates a situation where I am avoiding fantasy football roster exposure.
Follow Cynthia Frelund on Twitter.