Dr. Bob says: Lean Alabama in the title game, but it’s easy to see how Clemson wins

Bob Stoll of Dr. Bob Sports is one of the most seasoned, successful sports handicappers around. He was one of the first to use advanced analytics in sports handicapping after studying statistics and finance at U.C. Berkeley. You can find his historical college football performance here, and you can subscribe to his service here.

His Strong Opinions have a success rate of about 53.7% since 1999. He shared some of his lengthy, from-every-angle game previews here this season. Here are his thoughts on the national title game.

---

National Championship Game

Monday, January 9 – 5:30 pm Pacific

Alabama (-6 ½) 29 Clemson 20

Lean Alabama (-6 ½)

This is a rematch of last year’s exciting National Championship game that was won by Alabama 45-40. My math model correctly picked that game as closer than the spread and I thought Clemson had a really good chance to win that game straight up –- which they probably would have if not for a successful onside kick by Alabama that gave them the extra possession they needed to win.

Photo by Sean M. Haffey/Getty Images

This year the math favors Alabama even if I assume that Clemson’s offense will play relatively better in this game than their season rating and that Alabama’s pass attack will be relatively worse.

The Tigers do have hope for an upset, however, as they can surely win this game if they can shut down the Alabama ground game and force Bama’s unpolished freshman quarterback Jalen Hurts to throw the ball in more obvious passing situations, which is where he has struggled this season.

Clemson Offense vs. Alabama Defense

Deshaun Watson and the Clemson offense isn’t as good this season as it was last season but Watson still tends to raise his level of play against better defensive teams and throwing for an average of 6.4 yards per pass play against Ohio State is impressive given that the Buckeyes would allow just 4.2 yards per pass play to an average quarterback.

The level of the opposing defense has very little effect on Watson’s performance, as he averages close to the same yards per pass play against good defenses as he has overall during his career, which is the reason I picked Clemson to cover against Bama last season.

This season Watson has averaged 7.2 yppp against the 5 best pass defenses that he faced this season (Auburn, Florida State, Wake Forest, Virginia Tech, and Ohio State), which is only 0.4 yppp less than his overall average and is 2.1 yppp better than the 5.1 yppp than average quarterback would average against that same collection of opponents (compared to an overall rating of +1.6 yppp).

Watson was also relatively better against good defensive teams last season and he has now averaged 6.9 yppp in Clemson’s three playoff games – all against elite defensive teams (Oklahoma and Alabama last year), who would allow just 4.2 yppp to an average quarterback.

The Clemson rushing attack isn’t close to as good as it was last season, as the Tigers have averaged only 4.8 yards per rushing play (against teams that would allow 4.9 yprp). However, Clemson actually averaged more yards per rushing play against the better run defenses that they faced this season, which is due to Watson running the ball more often when he needs to run the ball more often, which not coincidentally is against good teams.

Watson didn’t run nearly as much this season as he did last season and he doesn’t need to run to beat a mediocre or bad team. But, he did run more in close games and I expect him to run more in this game, which should boost the Tigers’ rushing numbers even though Watson hasn’t been as good running this season as he was last season.

Last season Watson averaged 6.2 yards per run (1171 yards on 190 runs) and this season he’s averaged just 5.0 yards per run (635 yards on 128 runs). Alabama did a decent job defending Watson’s running in last year’s championship game, holding him to 4.9 yards per run, but Watson ran 18 times in that game and I think Watson will run the ball more than usual in this game.

Photo by Kevin C. Cox/Getty Images

Alabama has the best defense in the nation, allowing just 11.4 points and 248 yards per game at 3.9 yards per play (against teams that average 6.3 yppl against an average defense). This season’s defense is 0.5 yards per play better than last season’s elite stop unit and they are eager to get another shot at defending Watson after feeling embarrassed about how they played last year in this game.

I had some questions about the Crimson Tide pass defense heading into last week, but they shut down a very good Washington pass attack, allowing just 2.6 yards per pass play.

Overall the Tide defense has allowed just 4.3 yards per pass play this season to quarterbacks that would combine to average 7.1 yppp against an average defensive team.

The only quarterback to have a great game against them was Chad Kelly of Ole’ Miss, who averaged 9.4 yards on 43 pass plays in leading the Rebels to 43 points. Arkansas’ Austin Allen averaged 6.5 yppp (1.6 yppp less than his rating) in leading the Hogs to 30 points, so it is certainly possible for a good quarterback to have some success against Alabama’s secondary. However, Alabama also shut down some really good pass attacks –- limiting Western Kentucky to 3.9 yppp, Tennessee to 2.8 yppp, and Washington to just 2.6 yppp last week. Overall, Alabama has been relatively the same against good quarterbacks as they were against weaker ones they faced.

Alabama’s run defense is the best in the nation, as the Tide have yielded just 3.5 yards per rushing play to teams that would average 5.8 yprp against an average defensive team.

Clemson RB Wayne Gallman averaged a mediocre 5.1 ypr this season and he will have no chance of running the ball consistently against Alabama, which was also the case last year (just 45 yard on 14 runs against Bama). However, running quarterbacks are not impacted as much by a great run defense and Alabama gave up 5.3 yards per run to the 5 teams they faced with running quarterbacks this season (Miss, Tenn, A&M, Miss St, and Franklin of Auburn).

Kim Klement-USA TODAY Sports

I’ll assume Watson will run more often than normal in this game, as he did in other close games, and that Alabama’s strong run defense will have little affect on his running average. I still project just 3.4 yards per rushing play for Clemson (instead of 2.9 as the math model would project) since Gallman is expected to average just 2.8 ypr. I have also boosted the math projected yards per pass play for Clemson based on Watson’s history of playing relatively better against better defensive teams. Rather than 4.8 yppp that the math model would project using straight math, I project 5.7 yppp for Watson in this game.

Overall, I predict 348 yards at 4.7 yppl for Clemson in this game but Watson is projected to throw 1.55 interceptions, as he has a tendency to make dangerous throws, as evidenced by his 17 interceptions this season (2 last week).

Alabama Offense vs Clemson Defense

Alabama may be averaging 39.4 points per game but their offense contributed a more modest 31.9 points per game. Bama now has an incredible 15 non-offensive touchdowns this season, 4 on special teams and 11 on defense after adding one last week, but I still don’t think it’s likely that the Tide will score a non-offensive touchdown in this game, as Clemson has allowed zero special teams touchdowns and allowed an opponent to score on a fumble or interception return just once.

Of course, Washington had allowed zero non-offensive touchdowns until last week, so you never know. And, it’s actually more likely than last week that Bama can score on defense since Watson is an interception-prone quarterback and Alabama has returned 6 of their 16 interceptions for touchdowns.

As explained last week, Alabama would have scored a couple more points per game with their offense if they didn’t have all those non-offensive touchdowns since the offense would have had more possessions, so Bama still would have scored a good number of points even without the variance in non-offensive touchdowns.

Brett Davis-USA TODAY Sports

Alabama has averaged 466 yards and 6.6 yards per play this season under the leadership of freshman quarterback Jalen Hurts. Hurts is not a particularly great passer (I rate him at just 0.3 yards per pass play better than average) and he’s completed just 27% of his passes when under pressure this season, which is worse than average. Hurts was sacked 3 times by a very strong Washington defense last week and he averaged a pathetic 2.2 yards per pass play in that game.

Clemson has one of the best pass rushing defensive lines in the nation and I don’t think Hurts will have consistent success against that pressure. A regression equation to predict Hurts’ yards per pass play as a function of the rating of the opposing pass defense reveals that he is relatively less effective against better defensive teams. The equation has a slope of 1.30 and anything above 1.0 means that the opposing defense has more of an impact on performance than expected. Watson’s equation has a slope of 0.24, which is just the opposite.

As I did with Watson, I’ve decided to use the regression equation to project the yards per pass play for Hurts against a Clemson defense that is 1.7 yards per pass play better than average.

The regression equation projects just 4.6 yppp for Alabama, which is 0.2 yppp lower than what would otherwise be predicted.

Jason Getz-USA TODAY Sports

While Hurts isn’t likely to help the Bama offense with his arm he can still make plays with his legs (1043 yards at 6.5 yards per run) and has an excellent stable of running backs to hand the ball off to. Bama has averaged 265 rushing yards per game at 6.5 yprp (against teams that would allow 4.9 yprp) and the Tide ran for 297 yards at 6.5 yprp against a very good Washington run defense.

This is the area in which Alabama will need to shine to win this game and the Crimson Tide should have good success running against a Clemson defense that is good, but not great, defending the run. The Tigers are only 0.4 yards per rushing play better than average, allowing 4.8 yprp to teams that would combine to average 5.2 yprp against an average defense (adjusted for facing Auburn without star RB Pettway and facing Wake Forest with non-running QB Kearns rather than Wolford, who ran for 720 yards at 7.4 yprp this season).

Clemson’s run defense rating is just 0.1 yprp better than the average of the teams that Alabama faced this season and the Crimson Tide should get very close to their season average of 6.5 yprp (6.3 yprp is projected).

Overall, the math projects 262 rushing yards, 134 passing yards and 394 total yards at 5.6 yards per play for Alabama in this game. The 29 points that the math projects for Alabama may seem like a lot give how well the Clemson defense played in shutting out Ohio State last week. However, the Tigers did have some issues against other good offensive teams this season, allowing 34 points or more in each game, and an average of 37 points and 6.2 yards per play, to the other 4 good offensive teams they faced this season (Louisville, Florida State, Pitt, and Virginia Tech. I left out Auburn for reasons explained last week).

When the Ohio State game is included with those 4 other games against good offensive teams Clemson’s defense rated at 1.0 yards per play better than average, which is close to their season rating, so I’ll assume they play at their normal level in this game.

Summary

I think this game comes down to whether Clemson can defend Alabama’s rushing attack. If they do then they can unleash their pass rush and put more pressure on Jalen Hurts in obvious passing situations where he is likely to struggle.

However, Clemson’s run defense is not that strong and the Tigers allowed a combined average of 200 rushing yards at 6.6 yards per rushing play to the 4 good running teams that they faced this season (Louisville, Florida State, Pitt, and Ohio State). I excluded Georgia Tech because defending the option is not the same as defending a power rushing attack, and I excluded Auburn, who was actually a worse than average running team in the 4 games they played without star RB Pettway, who didn’t play against Clemson.

My math model favors Alabama by 9 ½ points (and 48 ½ total points) even after adjusting Clemson’s offensive ratings upward to account for Watson’s running and his tendency to play relatively better against better defensive teams, and downgrading Alabama’s pass attack to reflect Hurts’ relatively worse rating against good pass defenses.

I can certainly see a scenario in which Clemson wins this game but there are more sound reasons to back Alabama and I’ll lean with the Tide at -7 points or less. There is not enough value on the under to recommend a lean (i.e. the under has less than a 52.4% chance of winning), so I have no opinion on the total.