Analytics, Basic Stats, and Recent Historic Context: The ABC’s of Penn State Football – Aidan O’Connell vs. Sean Clifford – Relative Performance Index (RPI) Comparison
Sponsor: FTB’s Donors Club – the most direct way to support our efforts – is back for another year! (Sally Struthers voice) For $9.99 you can feed a starving blogger and get a cool FTB Koozie in return. JOIN HERE.
*Please remember to click the ‘Share My Address With For The Blogy’ box when checking out so we know where to mail your gift!
Introduction
Finally, the season is upon us! Bring on the first college football Saturday Thursday!
Much like last year, the Big Ten schedule makers left Penn State no room for non-conference creampuff foreplay as the Nittany Lions are forced to go in raw against another winning Big Ten West foe to get things started. And much like last year, one of the largest preseason questions marks looming over Penn State’s 2022 fortunes is the potential potency of the offense.
Rightly or wrongly, for better or worse, most of the uncertainty falls on the shoulders of QB1 Sean Clifford.
Clifford’s QB counterpart in the season opener is Purdue’s Aidan O’Connell – an unheralded walk-on and one time 8th string QB on the Boilermakers depth chart (seriously, that isn’t a joke. He was.) who somehow morphed into an All-Big Ten thrower. Statistically speaking, O’Connell ranked 2nd or 3rd in most major passing categories last year, only trailing behind Ohio State QB/G Wagon driver CJ Stroud, a Heisman finalist.
Now that we’ve set both QBs up, let’s compared the two, shall we?
In today’s analysis we will bring back the “Relative Performance Index” (RPI) that we used last year to predict the performances of Bo Nix and Sean Clifford in the White Out and the Iowa-Penn State game. As a brief primer, in the graphs below you’ll see the QB’s performance on a relative (percentage) basis versus the average performance the opposing defense allows against FBS offenses per game. For example, against Wisconsin in 2021 Clifford threw for 7.5 yards-per-attempt (247 yards on 33 attempts). Compare this to Wisconsin’s season average of 6.5 YPA allowed and you get an RPI for Clifford of 115% — meaning he exceeded projected expectations that game, which is obviously good. In this comparison, we ran analysis for all career games where Clifford and O’Connell faced FBS opponents AND played the whole game (i.e. I threw out mop up duty for Clifford in the 2018 Pitt and Kent State games and the Rutgers games in 2021 where he only played in the 1st quarter because of the old school flu).
Completion Percentage
We start with completion percentage, a stat where O’Connell (71.8%, 2nd in B10) far exceeded Clifford (60.2%, 6th) in 2021. But again, the pure A versus B comparison likely doesn’t tell the whole story here. The graph below shows the RPI in this metric for Clifford (left) and O’Connell (right) by Home (H) or Away/Neutral (A/N) splits color-coded by year. Note: this will be the format used in the next several graphs. The labeled line is the average RPI of each QB by the split and the boxes are 95% confidence intervals (more on that later).
So what stands out? Well, it’s interesting that location doesn’t seem to help or hamper either QB’s completion percentage. Both Clifford and O’Connell are consistent no matter where they play, with Clifford basically completing the same percentage of passes that opposing defenses allow on average (100% RPI). O’Connell, though, completes passes at a considerably higher rate than defenses typically allow (115% RPI). Oh, BTW, can you guess what 2021 game that ugly red dots anchoring the bottom of Clifford’s chart represents? If you guessed the Outback Bowl vs. Arkansas, gold star for you. In that forgettable display, Clifford connected with receivers on just 44% of his throws. O’Connell was very good at home in 2021 with several games of over 130% RPI (Wisconsin and Indiana).
Yards-Per-Attempt (YPA)
Next, we look at YPA. Here, O’Connell ranked 3rd in the B10 (8.4 YPA). Clifford ranked 10th (6.7 YPA)…so there’s a big disparity between the two. Much like completion percentage RPI, O’Connell output doesn’t move much in home/away splits. A couple of things can be inferred from this. One, he is a very mature and consistent quarterback who doesn’t get rattled by hostile crowds. In fact, some of his away performances were his best on the RPI scale. Also, maybe Ross-Ade Stadium doesn’t provide much home-field advantage (fingers crossed).
For Clifford we see a noticeable home/away split. Inside the friendly confines of Beaver Stadium, Clifford averages 112% RPI in YPA compared to 97% away. And he did that in spite of three hard-to-swallow homes performances in YPA last year – against Illinois, Indiana, and Michigan, Clifford’s RPI in YPA dipped below 80%. In 2021, Clifford had some very solid away performances in this metric, including at Wisconsin and Michigan State.
Passing Touchdowns and Pass Attempts
Aidan O’Connell is going to pass for more touchdowns than the expected defensive allowance. For his career, he averages 2.3 passing TDs per game. This is 0.8 TD above the allowed average for this dataset. Sean Clifford, in contrast, throws for roughly the expected RPI in TD passes when on the road, but gets a boost of 40% at home.
The 164-167% for O’Connell is impressive, but keep in mind that metric might be a tad inflated since Purdue only had 7 rushing TDs in 2021 compared to 35 passing TDs. Purdue likes to throw the ball, therefore, it’s not a shocker that the vast majority of the Boilers TD come via the air. Comparatively, Penn State’s offense generated 21 passing TDs and 10 rushing TDs against FBS competition in 2021. Despite the fact that these numbers are fairly anemic, it’s worth comparing the ratio versus Purdue.
Stats show the Boilermakers want to throw the ball about 20-30% more than an average offense. There have only been 3 games where O’Connell had less than 100% RPI in passing attempts and one of those was a 49-0 laugher against UConn in which Purdue scored 35 points in the first half.
Sean Clifford on the other hand has roughly the expected number of passing attempts on average. In 2021, his Pass Attempts RPI went up considerably (red dots), a shift that coincides with the arrival of Mike Yurcich along with a subpar run game. It’s hard to know how much of this was by design versus how much was because of the sputtering run game. I guess we’ll find out soon enough…
QB Rating
We conclude with QB Rating. Again, this metric – like the others we’ve included – seems to justify all of the preseason Aidan O’Connell hype. I firmly believe that Sean Clifford is also a very good quarterback, too, despite some struggles (especially after Iowa) in 2021. Clifford does tend to be better at home compared to O’Connell who has the same QB Rating average home and away. Interestingly though, for O’Connell, the bunching of his performances at home are very tight. Assuming that trend continues, he likely will not have a “bad” home game this Thursday. Clifford has shown more variability in his away performances.
The Penn State and Purdue Defense/Predicting the QB Performance
Of course, these RPI, projections aren’t static equations. The opposing defense has a big say in terms of what the allowances are for the RPI above and that can’t be forgotten. That is, in terms of absolute stats, a 150% RPI against Georgia or Alabama looks different than a 150% performance against UMass or UConn. So let’s look at the average performances of the Penn State and Purdue’s defenses over the last three years.
This is likely obvious, but the Penn State defense has been better than Purdue over the last 3 years. What might we expect from the QBs facing these defenses then? Below we take the Confidence Intervals of the RPI and the defensive allowances over the last 3 years and give expected statistical ranges for Clifford and O’Connell.
Clifford goes into the game with more expected variability. So we’re not completely sure what we’ll get, whereas O’Connell expected statistical variance is more narrow and consistent. O’Connell’s ceiling is higher in every category except Expected QB Rating. Clifford’s floor is lower in every category, yet historical data shows he’s capable of rising to the occasion ad outdueling O’Connell.
Can Aidan O’Connell take another step forward and cement himself as a topflight quarterback nationally? Will Sean Clifford start his sixth and final year with a bang? ¯_(ツ)_/¯
While what we just did was an interesting exercise, remember, the outcome of this game will hinge on much more than the arms of these two quarterbacks. With a new DC in Happy Valley and the Nittany Lions offense entering the 2nd year of Mike Yurcich’s system, along with some high impact transfers, Penn State is not the same team it was in 2021 – whether that be better or worse. Likewise, Purdue needs to replace some important weapons, an impact defensive player, and needs to heal up following a string of bad injury luck in camp.
Leave a Comment