BYU PER vs. OPP avg.

I realize that 99% of the people here (myself included) are over the Heaps vs. Riley debate. This isn't really about comparing the two quarterbacks as much as it is an attempt to determine how much of our QB improvement can be attributed to a decrease in opponent quality, and how much of it can be attributed to Riley's play.



If a QB is playing against a weak defense, he should be able to put up big numbers. Likewise, a good defense will make a QB struggle. Thus, I believe a more reliable metric for determining the quality of a QB's play is determining how that QB played in comparison to other QBs facing the same defense.



Jake Heaps PER Diff. vs. Ole Miss = -18.86

Opp. PER Avg. = 135.16

QBs faced = Jake Heaps (BYU #111), Larry Smith (Vandy #113), Aaron Murray (Ga. #30), Derek Carr (Fresno #62), AJ McCarron (Bama #31).

QB avg. PER rank = 58.



Jake Heaps PER Diff. vs. Texas = -5.20

Opp. PER Avg. = 103.71.

QBs faced = Taylor McHargue (Rice #72), Jake Heaps (BYU #111), Richard Brehaut (UCLA #52), Steele Jantz (ISU #93), Landry Jones (OU #18), Brandon Weeden (Ok. State #17).

QB. avg. PER rank = 60.



Jake Heaps PER Diff vs. Utah = -10.66

Opp. PER Avg. = 118.46

QBs faced = Matt Barkley (USC #19), Jake Heaps (BYU #111), Keith Price (Wash. #5), Brock Osweiler (ASU #42), Tino Sunseri (Pitt #86), Tino Sunseri's abysmal backup (not counting)

QB avg. PER rank = 53.



Jake Heaps PER Diff vs. UCF = -19.15

Opp. PER Avg. = 93.15

QBs faced = Chase Rettig (BC #92), Jake Heaps (BYU #111), Wesley Carrol (FIU #42), Rakeem Cato (Mar. #95), JJ McDermott (SMU #51).

QB avg. PER rank = 78.



Jake Heaps PER Diff vs. USU = -49.36

Riley Nelson PER Diff vs. USU = +75.64

Opp. PER Avg. = 129.36

QBs faced = Barrett Trotter (Auburn #71), Pete Thomas (CSU #80), Jake Heaps/Riley Nelson (#111/#13, thus given very unscientific #62), Brett Smith (Wyo #77), Derek Carr (#62)

QB avg. PER rank =70.



Riley Nelson PER Diff vs. SJSU = +25.96

Opp. PER Avg. = 133.64

QBs faced = Andrew Luck (Stanford #3), Richard Brehaut (UCLA #52), Tyler Lantrip (Nevada #67), Matt Christian (NMSU #12)*, Pete Thomas (CSU #80), Riley Nelson (BYU #13)*, Bryant Moniz (Hawaii #40).

QB avg. PER rank = 38.

QB avg. PER rank minus Christian and Nelson = 48.



Riley Nelson Per Diff vs. Oregon St. = +10.19

Opp. PER Avg. = 149.51

QBs faced = Russell Wilson (Wisc. #1), Richard Brehaut (UCLA #52), Brock Osweiler (ASU #42), Nick Foles (Ariz. #20), Riley Nelson (BYU #13)*

QB avg. PER rank = 26.

QB avg. PER rank minus Nelson = 29.



Note that stats do not include QB or defensive performances in games featuring FCS teams (i.e. Sac. State is not included in OSU's Opp. PER avg. or QB avg. PER rank).



Riley Nelson (and NMSU's Christian) was given a rank equal to what he would have with his current QB rating IF he was considered a starter for season stats purposes (he isn't). Despite being an unabashed Riley fanboy, I really doubt Riley's season PER would be in the 160s if he had been our starter from day 1. (I certainly doubt Matt Christian's PER would be so high if his only 3 opponents were Idaho, NM, and SJSU.) For that reason, in games where he appears, there's a separate QB average PER ranking with his rank omitted.



The bottom line is this - Riley Nelson has consistently performed above the average PER of each team he's faced, even in cases where (like OSU) the average included elite QB competition. There's no denying that our last three opponents have been worse against the pass than our first four - the key difference is that Heaps underperformed when compared to the average QB facing the same defenses, whereas Nelson is over-performing. That's where I think, statistically speaking, the "Riley's playing well because of weaker defenses" stops holding water.

Mars wrote:


Riley has been playing so well, it makes me wonder if we should actually recruit his little brother. Thoughts?

He committed to USU right before RIley go into the USU game. Not sure BYU had any interest, but I can see why he would have not pursued BYU given his style is similar to Riley.



Wondered the same thing. Maybe after his mission. Wouldn't that be too much? Imagine the whining from the land up north. IDK if he is a gamer like Riley - but it seems that he may well be.

DCR 36 - thanks for taking the time to put this together. I had posted elsewhere that we should look at this to deal with the SOS question.



Turns out that although there is some validity to the schedule being stronger - the performance difference is well beyond that factor alone.



Good analysis and good write-up.

Perfect analysis. The only way to get around it is to argue about the rest of the team not stepping up when Heaps plays, iow, blah blah blah.

Are those opponents defensive PERs, with or without the BYU game factored into that PER?

SpiffCoug wrote:
Are those opponents defensive PERs, with or without the BYU game factored into that PER?

With. Creates a feedback loop, but I was feeling lazy.

I can't argue against Riley outperforming Heaps, that is obvious. One question I do have though is how much better has our run game been with Riley as QB? How much of that improvement is because of Riley as QB and how much of Riley doing well as QB is due to the improved running game. I'm not sure there is a valid way to separate that out. When our running game improved last season is when we started seeing improvement in Heaps' PER.



[ Post made via Mobile Device ] http://cougarcorner.com/images/mobile.png" alt="" loading="lazy" style="max-width:100%">

It wasn't until the third game of BYU rushing for more than 200 yards did Heaps eclipse a 136 PER last year.

Most valid, statistically robust analysis to date. Well done.

Sheesh, if some of you guys didn't want hepas so much, maybe he could have saved the UT and U games. ;-)



[ Post made via Mobile Device ] http://cougarcorner.com/images/mobile.png" alt="" loading="lazy" style="max-width:100%">

Just a follow-up after reading a few comments: I don't think this is a truism in the sense that I don't think Heaps would ALWAYS throw below the median, or that Riley will complete this season without throwing below it (though I hope he does). I simply want to point out that, first, a shift occurred upon Riley's entrance as the starter, and second, that said shift could be attributable to something other than a decline in quality of opponent defenses.



Also, while 10 quarters of performance is enough (IMO) to show that Riley can "play," it still is a small sample size. Every QB has a bad game here and there, while Riley has done some stupid things (mostly the 2nd half vs. SJSU), he's yet to have a bad game. That could happen, but if it does, it would be more "Riley coming down to Earth" than "he finally played a decent defense."

Dcr36 wrote:
Just a follow-up after reading a few comments: I don't think this is a truism in the sense that I don't think Heaps would ALWAYS throw below the median, or that Riley will complete this season without throwing below it (though I hope he does). I simply want to point out that, first, a shift occurred upon Riley's entrance as the starter, and second, that said shift could be attributable to something other than a decline in quality of opponent defenses.



Also, while 10 quarters of performance is enough (IMO) to show that Riley can "play," it still is a small sample size. Every QB has a bad game here and there, while Riley has done some stupid things (mostly the 2nd half vs. SJSU), he's yet to have a bad game. That could happen, but if it does, it would be more "Riley coming down to Earth" than "he finally played a decent defense."

Yes - It doesn't guarantee anything. But, It is a useful model to predict what is more likely or less likely - but the all other things held equal part of modeling certainly leaves room for exception. But it is a strong predictor from a practical viewpoint because it has not only been true on the

average - it has been true every time so far.



So, if you get bored, or want to increase your sample size - every week, these teams play another team other than the occasional bye week. Add that data each week. Each week we play someone else and they come having already played several opponents. It will be interesting to see how many pluses click before we hit the first exception to the Riley rule.



Or add detail and list how many individual QBs bettered Riley's PER for a given opponent. Or has he outplayed every single one of them? He does definitely seem to be a D1 QB.



We could have some fun with this database to see who he has outplayed and who has outplayed him Vs common opponents. And yes - I recognize the imperfections due to different supporting casts, home Vs away, injuries and other changes to BYU, opponents, and opponents opponents team make-up and situation. None-the-less the correlation is too strong to dismiss - especially considering that Riley also outplayed Jake Vs UW, AFA, FSU before injury, and USU when they both played the same team the same day.



[ Post made via Mobile Device ] http://cougarcorner.com/images/mobile.png" alt="" loading="lazy" style="max-width:100%">

dmcougar wrote:
Most valid, statistically robust analysis to date. Well done.

Sheesh, if some of you guys didn't want hepas so much, maybe he could have saved the UT and U games. ;-)



[ Post made via Mobile Device ] http://cougarcorner.com/images/mobile.png" alt="" loading="lazy" style="max-width:100%">

Well - since you sort of brought it up. Heaps likely was below avg Texas Defensive PER in the first half and way below it in the 2nd half as has been true on a full-game basis with every opponent.



Riley exceeds the defensive PER of every opponent so far. That huge swing would have to be worth more than 2 points. He probably would have put them away in the first half and played well enough to keep our defense rested in the 2nd half.



Now - Vs Utah. Our defense should have kept pounding Wynn. We injured his shoulder but he

came back in and torched us only to get it hit against UW in his next game and go out for the season. Barring that. It is hard to say if Riley would have rallied the entire team to play enough better to not have the turnovers or overcome them. And U defense is - legit.



[ Post made via Mobile Device ] http://cougarcorner.com/images/mobile.png" alt="" loading="lazy" style="max-width:100%">

Nice analysis. Good insights.