College Bowl Games
Thu, Dec 26 1:00 PM PT
Rotation: 223, Odds: Miami Fla -6, Total: 50
Game Analysis view matchup stats
Lean – Miami-Florida (-6) 31 Louisiana Tech 22
Thursday, December 26 – 1 pm Pacific
Miami has had another disappointing season and the Hurricanes will be missing some key players in this game – but I still favor them to win by more than a touchdown against an overrated Louisiana Tech team even with this game being played about an hour from the Bulldogs’ campus (I gave Louisiana Tech half of a home field advantage).
Louisiana Tech is not as good offensively as their raw numbers suggest, as the 6.3 yards per play and 34.0 points per game was against a schedule of teams that would allow 6.7 yppl and 36.4 points per game to an average FBS offense. The Bulldogs were actually just 0.2 yppl worse than average offensively when excluding the two late-season games that quarterback J’Mar Smith missed and they’re about average as a unit when taking into account how rarely Smith throws interception (just 4 picks this season and 19 in 3 full seasons).
Miami’s defense rates at 0.9 yppl better than average for the season (4.9 yppl allowed to teams that would combine to average 5.8 yppl against an average defense) but defensive ends Trevon Hill and Jonathan Garvin are sitting, out along with LB Michael Pinkney, to prepare for the NFL draft (I think they all would be better served playing well in the bowl game given their current low draft rankings). Those 3 combined for 128 tackles, 14.5 sacks, and 16.5 other tackles for loss, which aren’t incredible numbers for 3 supposed NFL-caliber players. LB Shaq Quarterman and Greg Fousseau (14 sacks) are the stars of this defense and they’ll both be playing rather than punking out on their teammates. But, the 3 selfish players not playing are worth a combined 0.4 yppl and 2 points to Miami’s defense based on my algorithm. I project just 331 yards at 5.0 yppl for Louisiana Tech in this game.
Miami’s offense was mediocre from a compensated yards per play perspective this season, averaging 5.6 yppl against teams that would allow 5.6 yppl to an average attack, and the Hurricanes are hurt by the absence of injured RB DeeJay Davis (115 runs at 6.0 ypr), who was considerably better than the rest of the backs. WR Jeff Thomas, who caught just 31 passes and averaged a modest 8.4 yards per target is also sitting out this game with an injured back (he’s also leaving early for the NFL, which he probably won’t make). He’s worth just 0.2 points. Miami’s quarterback for this game is up in the air, as coach Diaz opened up the competition at the start of bowl practices. Jarren Williams was 0.6 yards per pass play better than N’Kosi Perry even with all the sacks that Williams takes and former Ohio State backup Tate Martell is also now in the mix. Miami’s offense is anywhere from 0.4 yppl worse than average to 0.1 yppl worse than average, depending on who is at quarterback (I just used their overall numbers for the math, adjusted for Davis and Thomas being out).
Louisiana Tech’s defense, like their offense, looks pretty good based on raw stats, as the Bulldogs have allowed just 23.7 points per game on 5.3 yppl. However, they’ve also faced a schedule of teams that would average only 19.0 points and 4.8 yppl against an average FBS defense and now the Bulldogs are without their best defender, CB Amik Robertson. Robertson had 5 interceptions and broke up 16 other passes while registering 8 tackles for loss (excellent for a cornerback) and he’s worth 0.5 yards per pass play and about 1.8 points. Louisiana Tech’s pass defense was terrible with Robertson (6.1 yppp allowed to quarterbacks that would combine to average only 5.1 yppp against an average defense) and whichever Miami quarterback gets the nod should post big numbers in this game. A big part of Miami’s pass game issues was the poor play of their offensive line that has given up 4.3 sacks per game, but Louisiana Tech has a terrible pass rush and now they don’t have their one good defensive back. Miami’s speedy receivers should be open and the Hurricanes’ quarterback should have more time to find them. I really think Williams, who is far more accurate than Perry (62.5% to 54.8% completions) would have an especially good game if he gets the nod to start, as he’s less likely to miss open receivers. Louisiana Tech does have a better than average run defense and Miami is projected to average just 4.8 yprp with Dallas out, but the Hurricanes should move the ball well through the air and I project 438 total yards at 6.6 yards per play for Miami in this game.
Overall the math favors Miami by 11.9 points (52.9 total points) and the absence of Robertson from Louisiana Tech’s defense is worth about the same as all 3 Miami defenders that are sitting out. I’m tempted to make Miami a play here but the Hurricanes are constantly underperforming. This season the Hurricanes lost games straight up as favorites of 14 points (Virginia Tech), 18.5 points (Georgia Tech), 20.5 points (FIU) and 9 points (Duke) and they only beat Central Michigan by 5 points as a 30.5-point favorite (although CMU turned out to be pretty good). On the other hand, Miami was 5-1 ATS this season when they were not favored by more than 7 points so they can certainly play well when they’re interested. The question is how interested are they to be playing this game? I’m sure that the team would like to avoid a losing season but Louisiana Tech is not a brand name opponent that is likely to get their blood flowing. I do know that Louisiana Tech will be excited to be playing in their home state against a big name team and coach Skip Holtz is traditionally good as an underdog (18-8-1 ATS) and good in bowl games (7-3 ATS; 4-1 as a bowl dog). I’ll resist making Miami a play (I was considering them for a Strong Opinion) but I’ll certainly lean with Miami at -7 points or less, as they have far more talent than an overrated Louisiana Tech team that played a very easy schedule and is without their most impactful defender.
Thu, Dec 26 5:00 PM PT
Rotation: 225, Odds: Eastern Mich +11.5, Total: 49
Game Analysis view matchup stats
Pittsburgh (-11.5) 31 Eastern Michigan 20
This comes down to how much the big favorite really cares about this game low-tier bowl game against a MAC opponent that they likely don’t respect. If Pitt plays at their normal level then they should win by 2 touchdowns or more but pre-New Year favorites of more than 7 points in non-major bowls are just 51-92-3 ATS and Pitt applies to an 8-56-2 ATS subset of that trend. Eastern Michigan, meanwhile, is 21-4-2 ATS as an underdog or pick since September of 2016, including 2-0 in bowl games.
Pittsburgh averaged just 20.4 points per game against FBS opponents while averaging 5.1 yards per play against teams that would allow 5.7 yppl to an average team. However, Eastern Michigan is horrible defensively, as the Eagles allowed 6.1 yppl to a schedule of teams that would average only 5.2 yppl against an average team. Eastern Michigan allowed an average of 30 points to offenses that rate the same as the Pitt attack and the math projects the Panthers to score 34 points on 487 yards at 6.5 yppl in perfect dome conditions. Pitt did under-perform their statistics this season by averaging just 4.1 yards per redzone opportunity, which is really low, but the redzone efficiency should improve in this game against an Eastern Michigan defense that was terrible in redzone defense (5.5 points per RZ allowed). Pitt receivers Mack and Ffrench are both listed as questionable for this game but it really doesn’t matter, as they were both below the rest of the team in yards per target and the pass rating was the same the last 3 games in which one or the other missed.
Eastern Michigan has a decent offense that averaged 29 points and 6.2 yards per play (against teams that would allow 6.3 yppl to an average team) but the Eagles averaged just 20.3 points in 4 games against average or better defensive teams Kentucky, Illinois, Central Michigan, and Buffalo and Pitt’s defense is much better than any of those teams. Those 4 teams combine to be just 0.2 yards per play better than average defensively while Pitt’s defense is 1.3 yppl better than average. The Panthers allowed just 4.7 yppl and 22.5 points per game against FBS teams that would combine to average 6.0 yppl against an average defense and my math projects just 296 yards at 4.6 yppl and 19 points for Eastern Michigan.
Overall the math favors Pitt by 15.4 points with 53.2 total points, which is adjusted for playing in a dome. However, as mentioned above, big favorites in minor bowl games have historically suffered a letdown and Eastern Michigan has been the nation’s best underdog bet the last 4 seasons. I’ll pass.
Fri, Dec 27 9:00 AM PT
Rotation: 227, Odds: Temple +4.5, Total: 53
Game Analysis view matchup stats
Lean – North Carolina (-6) 32 Temple 23
Friday, December 27 – 9 am Pacific
North Carolina needed to win their final two games to get to this bowl and the Tarheels are reportedly preparing for Temple with enthusiasm, as is generally the case for teams that had to win their finale to qualify for a bowl (such teams are 38-18 ATS if not facing another 6-6 team off a win). UNC was certainly better than their record suggests, as all 6 of their losses were by 7 points or less, including a 1-point loss to unbeaten #3 Clemson.
North Carolina moved the ball very well on offense, as they averaged 6.3 yards per play despite facing a schedule of teams that would allow just 5.3 yppl to an average attack. The Tarheels are led by quarterback Sam Howell, who averaged 7.3 yppp against a schedule of mostly good defensive teams that would allow 5.7 yppp to an average quarterback. Temple has a good defense, as the Owls are 0.9 yppl better than average on that side of the ball (4.9 yppl allowed to teams that would average 5.8 yppl against an average team) but UNC tends to control the ball and the math projects 454 total yards at 5.7 yards per play. I believe that North Carolina quarterback Sam Howell can outperform the projections against Temple in this game. Temple’s pass defense stats were very good overall (1.2 yards per pass play better than average) but the Owls gave up 8.1 yppp to the 3 really good passing teams that they faced this season (Memphis, SMU, and UCF). Those 3 teams would combine to average 8.3 yppp against an average defense, so Temple was just 0.2 yppp better than average against really good quarterbacks and Howell qualifies, as his compensated yppp rating is just 0.2 yppp worse than that of Memphis, SMU, and UCF quarterbacks.
Temple’s offense isn’t nearly as good as it was last season, as quarterback Anthony Russo doesn’t have the big play receivers he had in 2018. The Owls averaged only 22.4 points and 5.0 yards per play in their first 10 FBS games before exploiting defensively horrible Connecticut in their regular season finale and I rate that attack at 0.6 yppl worse than average. North Carolina is 0.2 yppl worse than average, allowing 5.8 yppl to teams that would average 5.6 yppl against an average defense, but that unit is better than the Temple offense and I project just 351 yards at 5.4 yppl for the Owls in this game.
Overall the math favors North Carolina by 8.5 points (with 54.9 total points) and I think there is upside in North Carolina’s offense given Temple’s struggles with good quarterbacks this season. For what it’s worth, Temple head coach Rod Carey is 0-6 straight up and 0-6 ATS in bowl games (UNC coach Mack Brown is 13-8 ATS). I’ll lean with North Carolina.
Fri, Dec 27 12:20 PM PT
Rotation: 229, Odds: Wake Forest +4, Total: 50.5
Game Analysis view matchup stats
Michigan State (-4/-3.5) 29 Wake Forest 24
Friday, December 27 – 12:20 pm Pacific
Michigan State had a very disappointing season and barely beat a dreadful Maryland team at home in their final regular season game to get to 6-6 just to qualify for a bowl game. The Spartans are a better than average team thanks to a defense that yielded just 22.7 points per game and 5.1 yppl despite facing a schedule of teams that would average 30.4 points and 6.0 yppl against an average defense. The defense is not as good now without suspended LB Joe Bachie, who was on his way to another All-Big 10 season before the PED suspension prior to week 11, but the Spartans are still good defensively and are certainly better than a mediocre Wake Forest attack that averaged 5.8 yppl against teams that would allow 5.8 yppl to an average offense. The Demon Deacons are higher scoring than their yards per play stats because they play at a fast pace and run a lot of plays. It looks like starting quarterback Jamie Newman will be ready to play, as he is listed as the #1 quarterback on the recently released depth chart. However, Newman was actually 0.3 yards per pass play worse than the team’s season average because while backup Sam Hartman was 1.0 yppp better than the team’s average compensated pass rating. Wake Forest lost leading receiver Sage Surratt after 9 games but freshman Donavon Greene has more than stepped up in this place, with 220 yards on 15 targets in 3 games and a 67% success rate, which leads the wide receiving group. The math projects Wake Forest with 367 yards at 4.9 yards per play but the Demon Deacons’ offense would likely surpass those numbers if Michigan State’s top cornerback Josiah Scott is unavailable to play. Scott is listed as questionable with a concussion but I’ll assume he’ll play.
Michigan State’s unimaginative offense was once again the problem with the Spartans, who averaged only 22.0 points and 5.3 yards per play against a schedule of teams that would allow 22.5 points and 5.4 yppl to an average team. Receiver Darrell Stewart may still be out (he missed the team’s last 4 games) but freshman Julian Barnett has matched Stewart’s mediocre yards per target number since he’s been out. I expect Michigan State to move the ball pretty well in this game, as Wake Forest’s defense is much worse than what the Spartans are used to seeing. The Demon Deacons allowed 29.3 points and 5.8 yppl in the regular season to teams that would combine to average 28.3 points and 5.7 yppl against an average defense and that unit is worse without top tackler LB Justin Strnad. Wake Forest’s run defense was 0.2 yards per rushing play better than average in 7 games with Strnad but they’ve been 0.4 yprp worse than average in 5 games without him (5.5 yprp allowed to teams that would average 5.1 yprp against an average team). The math projects 407 yards at 5.8 yppl for Michigan State in this game. Overall the math favors the Spartans by 5 points with a total of 53 points.
Fri, Dec 27 3:45 PM PT
Rotation: 231, Odds: Texas A&M -6.5, Total: 54
Game Analysis view matchup stats
Note: This game was released as a Best Bet before news of Oklahoma State’s star safety being out was announced on Thursday. It no longer qualifies at the current line and would not qualify now even if the line were still at +6.5 or +7. However, I still lean with Oklahoma State
Best Bet – *Oklahoma State (+7 at -115) 27 Texas A&M 29
Friday, December 27 – 3:45 pm Pacific
Texas A&M is the best 7-5 team in the nation, as all of their losses were to very good teams (Clemson, Auburn, Alabama, Georgia, and LSU). However, Oklahoma State is a good team too and the Aggies didn’t beat a team as good as the Cowboys.
Oklahoma State’s offense is very good, as the Cowboys averaged 6.6 yards per play against teams that would allow 5.5 yppl to an average team. Running back Chuba Hubbard is a 1st-Team All-American that ran for 1936 yards at 6.3 ypr and quarterback Spencer Sanders and backup Dru Brown combined for 7.3 yards per pass play (against teams that would allow 6.3 yppp to an average QB). Sanders’ thumb has healed and he’ll probably start this game but Brown is expected to play as well and both quarterbacks rated about the same in compensated yards per pass play, so it really doesn’t matter to me who’s behind center. Top receiver Tylan Wallace missed the final 4 games and remains out but Braydon Johnson and Jordan McCray have really stepped up in his absence and their 10.66 combined yards per target is exactly the same as Wallace’s 10.66 YPT this season. Oklahoma State’s pass attack actually rated at 0.2 yards per pass play better in the 4 games without Wallace so I felt that there was no reason to downgrade the offense. The Cowboys’ attack (1.1 yppl better than average) has an advantage over an A&M defense that has been 0.9 yppl better than average (5.4 yppl allowed to teams that would combine to average 6.3 yppl against an average team) but will be a bit worse without DT Justin Madubuike, who is sitting out to prepare for the NFL draft. The math projects 398 yards at 6.2 yppl for the Cowboys.
Texas A&M’s offense is nothing special, as the Aggies were 0.5 yppl better than average in the regular season (5.8 yppl against teams that would allow 5.3 yppl) and are worse than that heading into this bowl game with a depleted running back corps. The most notable absence is big-play back Cordarrian Richardson, who ran for 236 yards on just 25 runs. Taking those numbers out drops the A&M offensive rating to +0.4 yppl, which is the same as the Oklahoma State defense that yielded 5.8 yppl to teams that would combine to average 6.2 yppl against an average team. The Cowboys allowed 27.0 points per game and their opponent’s average rating is the same as Texas A&M’s offensive rating – so I don’t expect the Aggies to run away with this game with their offense. Unfortunately, it was announced on Thursday that Oklahoma State’s star safety Kolby Harvell-Peel would not play in this game due to an injured knee. That’s a huge loss for the for Cowboys, as Harvell-Peel is an elite safety that had 5 interceptions and broke up 13 other passes this season while ranking 3rd on the team in tackles. My algorithm projects his value at 0.4 yards per pass play and 1.5 points, which is a lot for a defensive player. The math now projects a mediocre 419 yards at 6.1 yppl for A&M in this game.
Oklahoma State is actually the better team from the line of scrimmage but A&M does have a 1.7 points edge in special teams and the math favors the Aggies by 2.6 points, with a total of 56.6 points. Oklahoma State was released as a 1-Star Best Bet at +6 points or more but would no longer qualify at the current line.
Fri, Dec 27 5:00 PM PT
Rotation: 233, Odds: Iowa -2, Total: 52
Game Analysis view matchup stats
Iowa (-2) 27 Southern Cal 26
Friday, December 27 – 5 pm Pacific
I feel that both of these teams are enthusiastic about playing in this bowl game and that should lead to a competitive battle between the offensively dominant USC team and a defensively stingy Iowa squad. USC’s offense blossomed in the first year under OC Graham Harrell’s version of the Air Raid offense, as the Trojans averaged 6.8 yards per play and 33.2 points per game against teams that would combine to allow 5.4 yppl and 23.2 points per game against an average attack. The Trojans are even better offensively when only taking into account the passing numbers of freshman Kedon Slovis, who took over in week 2 for a sub-par JT Daniels. Slovis completed 72% of his passes and averaged 8.1 yards per pass play (against teams that would allow 6.2 yppp to an average quarterback and USC’s offense rates at 1.5 yards per play better than average with Slovis at quarterback. Iowa’s defense should be up to the task of keeping the Trojans’ attack in check, as the Hawkeyes’ defense yielded just 13.2 points per game and 4.8 yppl to a schedule of teams that would combine to average 28.7 points and 5.9 yppl against an average defensive team. USC’s offense does have an advantage over the Iowa defense that allowed 6.5 yppl or more 3 times this season to good offensive teams Iowa State, Wisconsin and Minnesota (although they held Michigan and Penn State to a combined 4.4 yppl). The math projects 393 yards at 6.1 yppl for USC in this game.
USC also has a slight advantage when Iowa has the ball, as the Trojans rate at 0.4 yppl better than average defensively in 10 games with top defensive player Drake Jackson playing (0.3 yppl better than average in all games) while the Hawkeyes’ offense is just 0.3 yppl better than average. The math projects 383 yards at 5.6 yppl for the Hawkeyes.
While USC has the edge from the line of scrimmage, Iowa rates as a slightly better team overall due to their excellent specials teams. My math favors Iowa by 1.7 points with 54.7 total points in the good weather expected for this game but I’ll call for slightly lower total scoring given the 50.2 projected points based on a compensated (and adjusted) points model. I’ll pass.
Fri, Dec 27 7:15 PM PT
Rotation: 235, Odds: Washington St. +2.5, Total: 68.5
Game Analysis view matchup stats
Lean – Washington State (+2.5) 36 Air Force 34
Washington State is a good team with one major weakness that will not be fully exploited in this game. The Cougars have an elite offense but their inability to defend the pass killed them this season. That’s not going to be nearly as much of a problem against an Air Force option offense that runs the ball 85% of the time. Air Force will certainly have success when they do throw the ball, as quarterback Donald Hammond III averaged 12.3 yards per pass play (against teams that would allow 6.7 yppp to an average quarterback). However, Air Force should have a bit better than usual success running the ball, as Washington’s State’s run defense (0.5 yprp worse than average) is 0.3 yprp worse than that of the average defense that the Falcons faced this season. Air Force doesn’t usually throw the ball more unless their rushing attack is being slowed down so they’re not likely to take advantage of a Cougars’ pass defense that is 1.2 yppp worse than average (7.9 yppp allowed to quarterbacks that would combine to average 6.7 yppp against an average defense). If Air Force runs the ball 85% of the time, as they normally do, the Washington State’s defense would be about 4 points better than if facing an offense that runs and throws evenly – and that match up value is why my model favors Washington to win this game.
The math still projects 458 yards at 7.2 yards per play for Air Force in this game but Washington State’s prolific offense (514 yards per game and 7.1 yppl against teams that would allow 5.6 yppl to an average team) is scheduled to tally 480 yards at 7.6 yppl against an Air Force defense whose ability to defend the run (4.4 yprp allowed) is nearly wasted against a Cougars’ attack that will likely throw the ball 80% of the time. Air force is just average defending the pass (6.5 yppp allowed to quarterbacks that would average 6.5 yppp against an average defense), which is 0.3 yppp worse than the average defense that WSU faced in their 11 FBS games this season. So, there is also matchup value favoring Washington State when they have the ball because their pass-heavy offense plays away from the strength of the Air Force defense. Reports are that Washington State’s top receiver Brandon Arconado should be ready to go with his injured wrist (he’s listed as questionable), but the difference is only about a point if he doesn’t play.
Overall the math favors Washington State by 2.7 points, with a total of 70.9 points, assuming that the roof at Chase Field is closed. Air Force does apply to a 70-31-2 ATS bowl situation and the 3 service academies are a combined 34-14 ATS in bowl games since 1980, but I’ll still lean with Washington State, as it’s rare that you have this much match-up value in a game.
Sat, Dec 28 9:00 AM PT
Rotation: 239, Odds: Penn St. -7, Total: 60
Game Analysis view matchup stats
Lean – Memphis (+7) 28 Penn State 31
Memphis is extremely excited about playing in this major bowl game and the Group of 5 teams that make it to a major bowl have significantly outplayed expectations. In the 5 years of the current system the one Group of 5 team selected to play in a major bowl game are 3-1-1 ATS with the one spread loss being by just ½ a point. The average line is those games is +7.3 points and the G of 5 team’s average margin has been +2.6 points. Three of the five have won straight up as underdogs and the other two have lost by 8 points (as dogs of 8 and 7.5 points). Penn State may not be as thrilled to be playing in this game as they would be if they were playing a traditional national power and my math model doesn’t see the Nittany Lions as being that much better than Memphis.
The Memphis offense is very good, as the Tigers averaged 40.5 points on close to 500 yards per game at 7.2 yards per play against a schedule of teams that would combine to allow 30.3 points and 5.8 yppl to an average team. That offense will certainly be challenged by a very good Penn State defense that yielded just 14.7 points per game and 4.8 yppl against teams that would average 5.9 yppl against an average defense. That 14.7 points allowed is much lower than what would be projected by the Lions’ stats and they did allowed 31 points, 27 points and 28 points in 3 late season games against good offensive teams (Minnesota, Indiana and Ohio State), so Memphis should be able to score a decent numbers of points, especially given the enhanced field position provided by their elite special teams units that were 5.4 points better than average. The math projects 408 total yards at 5.9 yppl for the Tigers in this game.
Memphis has a pretty good defense that has been 0.4 yppl better than average (5.2 yppl allowed to teams that would average 5.6 yppl against an average team) but the Tigers’ stop unit is at a disadvantage against a very good Penn State attack that rates at 1.0 yppl better than average with quarterback Sean Clifford behind center. The math projects 445 yards at 6.5 yppl for Penn State in this game.
Overall the math favors the Nittany Lions by just 2.7 points, with 59.3 total points, and I’ll lean with Memphis plus the points.
Sat, Dec 28 9:00 AM PT
Rotation: 237, Odds: Iowa St. +3.5, Total: 54.5
Game Analysis view matchup stats
Best Bet – *Under (54.5) – Notre Dame 24 Iowa State 23
Lean – Iowa State (+3.5)
Both of these teams were really inconsistent offensively and the defenses for each team have the advantage in this game. Notre Dame averaged 6.3 yards per play but faced teams that would combine to allow 5.8 yppl to an average team averaged less than 6 yppl in 5 of their 12 games. Iowa State’s defense has a significant edge against the Irish attack, as the Cyclones yielded just 25.3 points and 5.3 yppl against teams that would combine to average 31.6 points and 6.3 yppl against an average defensive team. Notre Dame’s offensive rating is actually slightly less than the average rating of the offenses that Iowa State faced this season and my math projects just 5.15 yards per play for Notre Dame in what is expected to be inclement weather (likely rain with some wind). The Irish also had a tendency to play relatively worse against better defensive teams, as the slope of their yards per play as a function of their opponent’s compensated defensive yppl has a slope of 1.20. Anything above 1.0 indicates an offense that plays relatively better against weaker defenses and relatively worse against better defensive team. Ian Book was particularly inconsistent and built up his numbers with 13.6 yards per pass play in games against horrible defensive teams New Mexico and Bowling Green while mostly struggling against good pass defenses. Book averaged just 4.8 yards per pass play against teams that would 0.5 yppl better than average or more (Georgia, Virginia before losing All-American CB Hall, USC, Michigan, and Duke) and Iowa State’s pass defense rates at 1.1 yppl better than average (5.9 yppl allowed to quarterbacks that would combine to average 7.0 yppl against an average defense). Book’s slope predicting his yards per pass plays as a function of opponent’s pass defense is 1.43, so his relative performance level is particularly affected the quality of the opposing pass defense. In this case, Book would be projected to average only 5.32 yppp against Iowa State’s level of pass defense (including the weather adjustment), which is 0.45 yppp less than what the numbers would project if weighed all of Book’s pass plays evenly rather than putting more weight on his games against better defensive teams.
Iowa State’s Brock Purdy also has a couple of positive outliers against bad defensive teams that skewed his numbers up, as he averaged 13.8 yppp against ULM and Texas Tech. Purdy also had some good games against better than average defensive teams (7.8 yppp against Iowa, 9.7 yppp against TCU, 9.1 yppp against Oklahoma) but overall he was relatively worse against better defensive teams (although not the degree that Book is). Purdy is an accurate quarterback but his numbers are enhanced by a lot of big pass plays (5 games with a pass of at least 60 yards). However, Notre Dame’s defense allowed only 54% completions and did not allow a pass of over 50 yards all season. The Irish allowed just one quarterback (Jake Frohm from Georgia) to average more than 5.9 yppp against them, as they held the really good pass attacks of USC, Michigan, and Stanford (with Mills at QB) to a combined 5.7 yppp. Purdy’s regression equation as a function of opposing defense would project 5.25 yppp in this game and Notre Dame’s defensive regression equation yields a prediction of 5.28 yppp for the Cyclones in this game. Both are lower than the 5.44 yppp that would otherwise be predicted.
My math model liked the under here even before using the regression analysis to dampen the affect of the positive outliers against bad teams by each offense and now I think there is even more value. Part of the reason for the high line could based on the scoring of these teams, as the compensated points model would project 55.5 total points. However, both teams had very high combined redzone efficiency numbers, as Notre Dame’s games averaged 5.7 points per redzone opportunity (offense & defense combined) while Iowa State’s games averaged 5.4 points per RZ. The national average is 4.9 ppRZ and Notre Dame’s combined average is extremely high (I’ve never seen higher) and certainly unlikely to continue while Iowa State’s combined ppRZ is also due to regress towards expectations, as is the Cyclones’ 42.6% 3rd-down conversions allowed on defense, which is particularly high given how good their defense is overall.
My math model actually projects just 44.2 total points based on the projected stats but I decided to assume a higher points per redzone opportunity than my model would project. That gets the projected total up to 46.6 total points, which uses the overall scoring efficiencies of these two teams instead of the scoring efficiencies expected given the projected stats. Despite the unusually high combined points per redzone numbers, the Irish went under in 7 of 12 games and Iowa State also was under in 7 of 12 games in regulation (they went over vs Northern Iowa due to 29 OT points). If the redzone numbers regress some towards the mean then this game should go under with room to spare. I’ll go Under 54 points or higher in a 1-Star Best Bet.
While the math favors Notre Dame by 2.4 points, I’ll call for a slightly lower margin based on a 25-80-1 ATS bowl situation that applies to the Irish. Brian Kelly is also just 3-7 ATS in bowl games in his coaching career, including 0-3 ATS laying points. I’ll lean with Iowa State at +3 or more.
Sat, Dec 28 1:00 PM PT
Rotation: 241, Odds: LSU -13.5, Total: 76
Game Analysis view matchup stats
Best Bet – *Oklahoma (+14 at -115) 34 Louisiana State 42
Oklahoma is not getting the respect that they deserve. I understand that LSU is a great team but they were favored by only 7.5 points over Georgia in the SEC Championship game and the Bulldogs are not a better team than the Sooners – and certainly not 6 points better. LSU beat Georgia by 27 points while Oklahoma needed overtime to beat Baylor in the Big 12 Championship game. Those results have certainly influenced this line but Oklahoma actually out-played Baylor by 15.2 points from the line of scrimmage in that game, so they were better than they appeared and are not 14 points worse than any team – not even close.
Oklahoma coach Lincoln Riley has proven that his offense can score on any defense and the Sooners averaged 40 points in their 5 losses the last 3 years, including scoring 82 points against Georgia and Alabama in the playoffs the last two years against better defenses than this LSU defense. The Sooners have lost just one game by more than 7 points since early in the 2016 season and that was a spread covering 11-point loss against Alabama as a 14-point dog in last year’s playoffs. Alabama got up 28-0 in that game and Oklahoma scored 34 points in the final 3 quarters to make a game of it, so even a horrible start isn’t enough to keep the Sooners from staying within 14 points of a great team with the offense they’ve got.
Oklahoma’s offense is actually better than LSU’s offense from a compensated yards per play perspective, as Oklahoma has been 3.0 yards per play better than average with Jalen Hurts in the game at quarterback (8.5 yppl against teams that would allow 5.5 yppl to an average team) while LSU has been 2.7 yppl better than average with Joe Burrow in the game (8.1 yppl against teams that would allow 5.4 yppl). Oklahoma’s defensive rating, meanwhile, has only 0.2 yppl worse than LSU’s defense this season, rating at 0.8 yppl better than average (5.3 yppl allowed to teams that would average 6.1 yppl against an average team) while the Tigers’ defense was 1.0 yppl better than average with their starters in the game (5.1 yppl allowed to teams that would average 6.1 yppl against an average team).
For the season these teams are very even from the line of scrimmage (Oklahoma is actually slightly better), but LSU’s defense did improve later in the season while Oklahoma is missing two key starters in DE Ronnie Perkins (suspended) and S Delarrin Turner-Yell (collarbone injury). I rate LSU’s defense at 1.2 yppl better than average with their current lineup and dampening the affect of their negative outliers (allowing 8.5 yppl to Mississippi) while Oklahoma’s defense goes from 0.8 yppl better than average to just 0.5 yppl better than average with Perkins and Turner-Yell out. Still, even with those adjustments the math model would project a pretty close game.
Digging in deeper I found that Oklahoma’s offense was relatively worse against better defensive teams while LSU’s offense was relatively the same regardless of level of opposing stop unit. Using the regression equations to predict Oklahoma’s rushing and passing as a function of the level of the opposing run and pass defenses would still yield a prediction of 493 yards at 7.2 yppl for the Sooners in this game, which is includes a negative adjustment to their rushing attack with backup RB Rhamondre Stevenson out (suspended). LSU, meanwhile, is projected at 535 yards at 8.1 yppl against what is still an underrated Oklahoma defense – even without two starters.
Overall, the math favors LSU by just 6.6 points (with 78.2 total points in perfect dome conditions) based on the projected stats but Oklahoma does tend to underperform their stats – and I get LSU by 9.5 points after taking that trend fully into account. However, a part of that discrepancy between Oklahoma’s actual scoring margin and their stats-based projected scoring margin is likely random, so 7.5 or 8 points is a more reasonable projected margin for this game.
My projections assume that both starting running backs, Kennedy Brooks for Oklahoma and Clyde Edwards-Helaire for LSU will both be 100%. That’s probably true for Brooks, who teammates say looks great, but Edwards-Helaire has a hamstring injury and is being considered a game-time decision. LSU’s offense would really suffer without Edwards-Helaire, as the other 3 backs have combined for just 551 rushing yards at 4.7 ypr while Edwards-Helaire averaged 6.5 ypr this season. The difference in the math if he doesn’t play is 2.3 points and there is solid value on Oklahoma even if I assume that LSU’s star running back 100%, which is likely not the case even if he plays. I’ll take Oklahoma in a 1-Star Best Bet at +14 at -115 odds or better (Strong Opinion down to +13).