Since my brother-and-law is currently a student at Mizzou, last weekend my wife, some of my wife's family, and I (later joined by some of our friends from band and ΚΚΨ) headed down to Columbia for a football game. We expected Purdue to be tested by a team that has dominated the SEC East in recent history, but struggled last season.
Then something else happened.
My Purdue fandom has been dominated by falling short in big moments under Hope, then falling short on any moment under Hazell. (I'm in the Mechanical Engineering Class of 2015 - that should give a reference of what how little I know of successful Purdue teams.) I'm not sure how to react to something like this. Other than by being in awe of some of the numbers.
Declawing the Tigers
Purdue's defense was fantastic in Columbia on Saturday. The Tigers' offense, coming into the game, was not a bad group, and expectations were for a shootout. Instead, Mizzou managed only 203 yards of offense with an average of only 3.90 yards per play, compared to a national average of 5.94 yards per play. They managed only 0.23 points per drive. They passed for 133 yards on 28 attempts (with a 42.9% completion percentage), creating 4.75 yards per attempt (National Average: 7.62 yards/attempt) [1,2,3]. The advanced stats did not like Mizzou quarterback Drew Lock any more than the traditional stats; he had a QBR of 14.8 [4]. The Tigers rushed for 70 yards on 24 attempts, with 2.92 yards per attempt (National Average: 5.27 yards/rush) [2,4]. It was a dominant performance for the Boilers, especially compared to better performances by THE Ohio Bobcats against the Boilers. For the sake of any Mizzou fans reading this, I'm going to move on quickly from discussing this element of the game.
Tale of Two Halves
Purdue's offense continued its string of great performances in the first half, scoring efficiently with four points per drive and managing 6.74 yards per play. This allowed the Boilers to enter halftime with a 28-3 lead. The second half was much less productive, with a rather inefficient 1.17 points per drive and 2.90 points per play. Conservative play calling, as Brohm stated in his press conference on Monday, definitely played a major role in that drop off [5]. The stats bear that out: while the Boilers had only 0.55 plays more per drive in the first half, they managed over double the yards per play and almost four times the points per drive over the Tigers. Ending the day, the efficiency stats were alright (4.97 yards per play; 2.69 points per drive), with some nice (but not particularly useful) total stats (477 yards; 35 points) [1].
Individual offensive performances were also very good. Blough's QBR was an excellent 91.7, with a 78.6% completion percentage and 6.7 yards per attempt. Sindelar saw limited action but posted a near perfect 96.6 QBR on 6 pass attempts, with a 66.7% completion percentage and 14.2 yards per attempt. While Blough's much larger number of attempts (28 v. 6) show him cementing his starting role, Sindelar put in a much improved performance compared to previous weeks [4].
The running backs got a lot of work as well, as seen in Figure 3. The second half made the numbers less efficient than the national averages, but a number of backs still put in performances that got the job done. Also in Figure 3 is the receivers' performances; unlike in past weeks, the ball was spread around with more than just tight ends Hopkins and Herdman showing the ability to catch. Going forward, this is a major improvement.
Granted, Mizzou's defense has struggled mightily throughout the season - and not just against Purdue. They have given up 36.3 points per game (113th nationally), and 5.9 yards per play (National Average: 5.94). This included 8.38 yards per passing attempt (National Average: 7.62 yards/attempt). Using Bill Connelly's (of Football Outsiders and Football Study Hall) efficiency metrics, the Tigers looked even worse with a 44.7% defensive efficiency - ranking 102nd nationally [7]. Like the game against THE Ohio, this is a good season start to build on going forward.
The B-Word
After this game, there have been significant conversations regarding the possibility of a bowl game for Purdue. As I was researching for this article, I noticed over at Football Study Hall Bill Connelly had simulated the season using his S&P+ metric and created win distributions. I owe a lot of thanks to Connelly for a lot of the metrics he developed that I (and others) use extensively. Especially given how otherwise I'd have to take those metrics and simulate win distributions, I thank him for doing it for me. So, ripping off from him, here is how Purdue's win distribution looks going forward:
In an average simulation, Purdue's wins total 5.07, meaning a bowl trip is based on a lack of teams (and APR scores). In 42% of simulations, however, the Boilers have 6 or more wins, and punch a ticket to a bowl game. While I wouldn't recommend booking hotel rooms in Pasadena just yet, if you are booking flights anywhere remember Southwest will let you cancel flights and receive full credit for use on another flight (including LAX for those that dream, and Detroit for those who are slightly more realistic).
So You're Saying We Have A Chance...
In a probabilistic sense, you always have a chance to win any game at kickoff. It may not be a large chance, but it is a chance. In the run up to this season, I figured it was a chance along the lines of "roll a 20 to the find weakness of the giant monster, then roll a 20 again to attack the monster" what with Michigan being Michigan and Purdue being what Purdue had been under Hazell. I saw the schedule and thought the athletic department was crazy for scheduling homecoming against Michigan. Did they want another beat-down for homecoming? At least this time it wouldn't be from Illinois...
Then Brohm happened. These Boilers looked like a real, live, half decent football team. But Michigan is still Michigan. They rank 8th in both the AP and Coaches' Polls, and 11th in S&P+ [9,10]. This is a team that, in some ways, looks too imposing even for a re-energized Boilermaker squad to play well against. They rank second in the nation in defensive S&P+, expected to give up 12.8 points in a game against an average offense [10]. They are allowing opponents 5.03 yards per passing attempt, 2.3 yards per rush, and 3.4 yards per play [11]. Michigan's defense is very, very good; it will be a real test for Brohm's offense to play against them.
The Wolverines' offense, however, isn't quite as good. S&P+ ranks them 49th in the country, expected to score 32 points against an average team. In the passing game, they have managed only 7.80 yards per attempt (National Average: 7.62 yards/attempt). Bill Connelly's efficiency metrics like the Wolverines less than the more traditional stats: their success rate of 34.9% puts them at 99th in the nation, with the national average at 40% [12]. Their rushing attack has been less than stellar as well: 4.7 yards per rush (National Average: 5.27 yards/rush) and a 38.1% success rate (ranked 98th) [3,12]. Our defense, 75th in nation with an S&P+ of 29.9 (expected points per game) held in check a more potent Louisville (12th, 38.6) attack and embarrassed Mizzou (29th, 35.3) [10]. If Michigan continues to struggle and our defense can continue to play well, it gives Purdue an opportunity.
As calculated by Connelly, Purdue has a 15% chance to win, with a projected score of 34.9 to 17.2 [8]. That's not great (i.e. roll an 18 or above). Even though Purdue is still an underdog, stranger things have happened. At the very least, Purdue should make the game respectable for quite a while and, if Michigan has offensive struggles again, possibly have moments where the game will be close.
Questions, Comments, Concerns?
This is my third of these summaries. If you've read these, thank you. I am not a professional writer or football analyst. It help improve these posts, I would love feedback. Whether that is blind spots in my knowledge, a preference at looking at team or individual stats, criticism of tone, issues of length (i.e. should I split the summaries and preview section), or a desire to see topics other than football/sports looked at (I promise, more Fiction v. Engineering is coming), let me know.
Boiler Up!
References:
[1] http://www.espn.com/college-football/playbyplay?gameId=400933860
[2] https://www.sports-reference.com/cfb/boxscores/2017-09-16-missouri.html
[3] https://www.sports-reference.com/cfb/years/2017-team-offense.html
[4] http://www.espn.com/college-football/boxscore?gameId=400933860
[5] https://www.hammerandrails.com/2017/9/18/16329672/michigan-at-purdue-depth-charts-and-coach-brohm-speaks
[6] https://www.sports-reference.com/cfb/schools/missouri/2017.html
[7] https://www.footballstudyhall.com/pages/2017-missouri-advanced-statistical-profile
[8] https://www.footballstudyhall.com/pages/2017-purdue-advanced-statistical-profile
[9] http://www.espn.com/college-football/rankings
[10] http://www.footballoutsiders.com/stats/ncaa
[11] https://www.sports-reference.com/cfb/schools/michigan/2017.html
[12] https://www.footballstudyhall.com/pages/2017-michigan-advanced-statistical-profile
Then something else happened.
 |
| Figure 1: A Different Kind of Purdue Game |
Declawing the Tigers
Purdue's defense was fantastic in Columbia on Saturday. The Tigers' offense, coming into the game, was not a bad group, and expectations were for a shootout. Instead, Mizzou managed only 203 yards of offense with an average of only 3.90 yards per play, compared to a national average of 5.94 yards per play. They managed only 0.23 points per drive. They passed for 133 yards on 28 attempts (with a 42.9% completion percentage), creating 4.75 yards per attempt (National Average: 7.62 yards/attempt) [1,2,3]. The advanced stats did not like Mizzou quarterback Drew Lock any more than the traditional stats; he had a QBR of 14.8 [4]. The Tigers rushed for 70 yards on 24 attempts, with 2.92 yards per attempt (National Average: 5.27 yards/rush) [2,4]. It was a dominant performance for the Boilers, especially compared to better performances by THE Ohio Bobcats against the Boilers. For the sake of any Mizzou fans reading this, I'm going to move on quickly from discussing this element of the game.
Tale of Two Halves
 |
| Figure 2: Purdue Offensive Performance By Half [1] |
Individual offensive performances were also very good. Blough's QBR was an excellent 91.7, with a 78.6% completion percentage and 6.7 yards per attempt. Sindelar saw limited action but posted a near perfect 96.6 QBR on 6 pass attempts, with a 66.7% completion percentage and 14.2 yards per attempt. While Blough's much larger number of attempts (28 v. 6) show him cementing his starting role, Sindelar put in a much improved performance compared to previous weeks [4].
The running backs got a lot of work as well, as seen in Figure 3. The second half made the numbers less efficient than the national averages, but a number of backs still put in performances that got the job done. Also in Figure 3 is the receivers' performances; unlike in past weeks, the ball was spread around with more than just tight ends Hopkins and Herdman showing the ability to catch. Going forward, this is a major improvement.
 |
| Figure 3: Purdue Individual Rushing and Receiving Statistics [2] |
The B-Word
After this game, there have been significant conversations regarding the possibility of a bowl game for Purdue. As I was researching for this article, I noticed over at Football Study Hall Bill Connelly had simulated the season using his S&P+ metric and created win distributions. I owe a lot of thanks to Connelly for a lot of the metrics he developed that I (and others) use extensively. Especially given how otherwise I'd have to take those metrics and simulate win distributions, I thank him for doing it for me. So, ripping off from him, here is how Purdue's win distribution looks going forward:
 |
| Figure 4: Purdue Win Distribution [8] |
So You're Saying We Have A Chance...
In a probabilistic sense, you always have a chance to win any game at kickoff. It may not be a large chance, but it is a chance. In the run up to this season, I figured it was a chance along the lines of "roll a 20 to the find weakness of the giant monster, then roll a 20 again to attack the monster" what with Michigan being Michigan and Purdue being what Purdue had been under Hazell. I saw the schedule and thought the athletic department was crazy for scheduling homecoming against Michigan. Did they want another beat-down for homecoming? At least this time it wouldn't be from Illinois...
Then Brohm happened. These Boilers looked like a real, live, half decent football team. But Michigan is still Michigan. They rank 8th in both the AP and Coaches' Polls, and 11th in S&P+ [9,10]. This is a team that, in some ways, looks too imposing even for a re-energized Boilermaker squad to play well against. They rank second in the nation in defensive S&P+, expected to give up 12.8 points in a game against an average offense [10]. They are allowing opponents 5.03 yards per passing attempt, 2.3 yards per rush, and 3.4 yards per play [11]. Michigan's defense is very, very good; it will be a real test for Brohm's offense to play against them.
The Wolverines' offense, however, isn't quite as good. S&P+ ranks them 49th in the country, expected to score 32 points against an average team. In the passing game, they have managed only 7.80 yards per attempt (National Average: 7.62 yards/attempt). Bill Connelly's efficiency metrics like the Wolverines less than the more traditional stats: their success rate of 34.9% puts them at 99th in the nation, with the national average at 40% [12]. Their rushing attack has been less than stellar as well: 4.7 yards per rush (National Average: 5.27 yards/rush) and a 38.1% success rate (ranked 98th) [3,12]. Our defense, 75th in nation with an S&P+ of 29.9 (expected points per game) held in check a more potent Louisville (12th, 38.6) attack and embarrassed Mizzou (29th, 35.3) [10]. If Michigan continues to struggle and our defense can continue to play well, it gives Purdue an opportunity.
As calculated by Connelly, Purdue has a 15% chance to win, with a projected score of 34.9 to 17.2 [8]. That's not great (i.e. roll an 18 or above). Even though Purdue is still an underdog, stranger things have happened. At the very least, Purdue should make the game respectable for quite a while and, if Michigan has offensive struggles again, possibly have moments where the game will be close.
Questions, Comments, Concerns?
This is my third of these summaries. If you've read these, thank you. I am not a professional writer or football analyst. It help improve these posts, I would love feedback. Whether that is blind spots in my knowledge, a preference at looking at team or individual stats, criticism of tone, issues of length (i.e. should I split the summaries and preview section), or a desire to see topics other than football/sports looked at (I promise, more Fiction v. Engineering is coming), let me know.
Boiler Up!
References:
[1] http://www.espn.com/college-football/playbyplay?gameId=400933860
[2] https://www.sports-reference.com/cfb/boxscores/2017-09-16-missouri.html
[3] https://www.sports-reference.com/cfb/years/2017-team-offense.html
[4] http://www.espn.com/college-football/boxscore?gameId=400933860
[5] https://www.hammerandrails.com/2017/9/18/16329672/michigan-at-purdue-depth-charts-and-coach-brohm-speaks
[6] https://www.sports-reference.com/cfb/schools/missouri/2017.html
[7] https://www.footballstudyhall.com/pages/2017-missouri-advanced-statistical-profile
[8] https://www.footballstudyhall.com/pages/2017-purdue-advanced-statistical-profile
[9] http://www.espn.com/college-football/rankings
[10] http://www.footballoutsiders.com/stats/ncaa
[11] https://www.sports-reference.com/cfb/schools/michigan/2017.html
[12] https://www.footballstudyhall.com/pages/2017-michigan-advanced-statistical-profile
Love it. Keep it up.
ReplyDelete