Perhaps the BEST B1G Forum anywhere, here at College Football Fan Site, CFB51!!!

Now, allow me to turn it on its head and agree with @OrangeAfroMan 

I speak of variance as if it's bad, but variance can be either good or bad. In the NFL, where there's more parity and less chance of being excluded from the playoffs by a single loss, it makes sense to run a higher variance system if the rewards are greater. And from what I can see, "success rate" tends to be higher passing than rushing in the NFL, so it *should* be a more successful system overall.

So let's say you're an NFL team with 9-7 talent. Running a higher variance system [heavy passing] might cause you to lose a game you shouldn't, but because success rate is higher for passes than rushes, it might net you two or three additional wins relative to what you "should" have. So you have the opportunity to boost your record from 9-7 to 10-6 or 11-5, which might move you from wild card weekend to division winner, get you better HFA opportunities in the playoffs, etc.

Likewise let's say you're a college team with 7-5 or 8-4 talent. You're not likely to be in the hunt for the CFP. A single loss won't derail your season. But if you can rack up 2-3 wins over "average" expectation, you might be in a higher-tier bowl game, you might be playing in your CCG, etc. So it makes sense to play the risk-reward game.

The areas where it's harder to do this, though, is when you're a legitimate CFP contender. Variance can kill you, because a single loss (particularly a bad upset) can eliminate you from the CFP. 

In 2018, Ohio State came to West Lafayette. Dwayne Haskins threw 73 passes to only 25 team rushes. Still, Ohio State outgained Purdue 546 to 539. Ohio State had 31 first downs to Purdue's 24. Ohio State was 50% on 3rd down to Purdue's 40%. And Purdue won 49-20. 

The big difference in that game is that their variance came at the wrong time--they had 4 drives (IIRC) stall inside Purdue's 10. Those drives finished with only 6 points, as they had a missed field goal and turnover on downs.  On any other day those 4 drives against a team like Purdue would have probably led to 17-20 points (or more). Wisconsin would have done so if they'd gotten down there... 

imo, sacks should be a passing stat. rpo throw a wrinkle in it a little, but imo, it should count towards the type of play it was intended to be.

also, int's should count against wr/te/rb if it was a tipped ball or obviously catchable and they messed it up.

also, int's should count against wr/te/rb if it was a tipped ball or obviously catchable and they messed it up.

I get the idea, but it gets tough to determine.  What if they ran the wrong route?  Does it get charged to the OT if he whiffed his block and the QB got drilled as he released it?  What if the missed blitzer though, was the result of the QB misreading the rush.=?

"3 things can happen when you throw a pass, and 2 of them are bad."


- WW Hayes

I mean, a lot of things can happen on a run or a pass. 

According to Football Outsiders, the best OL in the league in 2019, Dallas, has a 13% likelihood of a run being stuffed, i.e. tackled at or behind the LOS. The worst, Miami, was 26%, but only 9 of 32 were 21% or higher.

Contrast that to completion percentage. Brees was at the high end, besting 75% this year, but only 3 players were over 70%. The bulk were between 60-70%, with only three players finishing below 60%. (Qualified by minimum 200 attempts--there are several starters who didn't qualify with 200 attempts below 60% though, most of whom failed to qualify due to seasons limited by injury).

So not taking into account fumbles, interceptions, sacks, somewhere under 20% of runs will be zero (or negative) yards for most teams, while somewhere between 30-40% of passes will be zero yards.

That's my point when it comes to variance. The NFL average is 4.26 ypc, so two "average" runs should leave 3rd and short. Two mediocre runs in a row, let's say half of average, or one stuffed run and one average run, and you might be facing 3rd and 5 or 6, where you're still in intermediate territory and may have a run or a pass option. Two incomplete passes in a row and you're in 3rd and 10 and in an obvious passing situation.

The way to win football games is to move the chains in sustained drives. Variance where 30-40% of your plays go for zero yards can stall drives even if the average of those plays is higher.

FWIW, that gap in College is 11.1 percent (Army) to 32.6 (Tennessee). Not sure if sacks are filtered out.

What I meant to get at earlier in a wishbone-offense example I cited, there was a very practical reason that there wasn't much passing.  The QBs often had banged-up shoulders from being tackled as runners.  If you can't complete passes unless the receiver is wide open and not too far downfield, it's not worth it to throw the ball as part of your normal offense.  If the QB, to avoid contact, routinely pitched the ball on the triple option, then the defenders would quickly learn that he was not a threat to run and ignore him.  And the HB getting the pitch would have two defenders in his face when he got the ball.

Oklahoma's best wishbone teams seldom threw the ball, but when they did it was usually either an incompletion or a big gain, because the WR or TE was wide open.  But they wouldn't have been wide-open had OU been throwing the ball 15-20 times a game instead of 4-6 times.

That's the difference between a pro-style offense (or WCO, or run and shoot) and a triple-option offense where the QB run is one of the options on nearly every play.  If the QB is a good passer, first of all, he won't go to a school that runs a triple-option offense, and second, he's probably not a good enough runner to make the QB-run option completely effective.

So the discussion about the pass/run split is different when you are talking about a triple-option team than it is when you are talking about an offense like Bama's or Clemson's or LSU's or Ohio State.  The QB run in those offenses may well be part of the game plan, but it's not designed into the vast majority of plays.

Very true - outlier offenses may have their own diminishing returns point.  Their bell curve may be off-set from the norm by 20% or so.  You have your Switzers on one end and your Leaches on the other and they may stray so far from the norm that they require their own math.

Have you analyzed Leach's offenses for pass/run balance?  It's my impression that he runs a lot more than people think he does, but I don't have stats to back that up.  And I'm talking about real runs, not swing passes to backs.