Explaining Points per Drive vs Points Per Game

Longestday

Helluva Engineer
Featured Member
Messages
2,856
Which is better if all scores are TDs:
1. An offense scores on 50% of its drives and averages 8 drives per game
2. An offense that scores 25% of the time and averages 16 drives per game

They are the same in points per game and have the same points per game stats of 28 points. But team 1 only gave the other team 8 times to score. If the other team cannot be as increadible efficient then the other team loses. 50% is pretty darn good! But 25% is not that great and if the other team had a higher scoring rate then team two loses. Same score with a good team (1) and a bad team (2).

Which is better if all scores are TDs?
1. An offense scores on 50% of its drives and averages 8 drives per game = 28 points
2. An offense that scores 50% of the time and averages 16 drives per game = 54 points (not the realm of normal scores)

They are much different in points per game (by 2X). They are the same ability. If team one had the same number of drives they would score the exact same number of points. Again the opposing offense would have to have an equal or better than 50% scoring per drive to beat team one (same as team 2). See how comparing points per game does not work?

Comparing defenses:
1. If the opposing team only has 8 drives and scores on all drives the max score would be 56
2. If the opposing team has 16 drives and scores on all drives the max score would be 112 points

You expect points to be skewed lower in games with fewer drives. You need to compare the actual score to the possible score the see if you did well. You don't compare your score in 8 drives to someone who had 16 drives to see how good you are. See how points per game does not work? to compare defense.

Defense effects the number of drives:
1. If the defense stops all 8 drives but allows 11 plays per drive (88 plays on D)
2. If the defense stops all 16 drives but only allows 3 plays per drive (48 plays on D)

Team 2 limits the time of the game the other team is on the field allowing their offense to have more drives and or time.

Suppose you have the following teams with the same 50% scores per drive all TD's and the opposing O scores 30% per drive.

1. Offense that plays an average of 70 snaps per game and 11 meaningful drives because they wait till the last second to hike the ball and run the ball a lot keeping the clock going
2. The same offense as above that runs 55 snaps per game and 8 meaningful drives because the defense cannot consistently stop 3rd downs allowing the other team to stay on the field.
3. Offense that plays an average of 90 snaps per game and averages 16 drives per game becauase they pass and hurry the snap.

Team 1 will score 38.5 points averaged versus the opposing team of 23.1 points
Team 2 will score 28 points averaged versus the opposing team of 16.8 points
team 3 will score 54 points averaged versus the opposing team of 33.6 points

This is the same game with different points scored. You cannot measure point per game and compare all games.

This is the same as allowing some games to be 1 hour and some games to be 2 hours. Would you compare scores between these two games. You would not because you would think that you can score more points in 2 hours than 1 hour. You would more than likely have more drives in 2 hours than 1 hour.

Throw out your points per game, yards per game, and most of your per game stats. Per drive and or per play stats will help normalize the games to help you understand the true differences in a teams effectiveness.

I would want a 50% effectiveness on O and lower points per game than a 30% effectiveness and more points per game. You win more with 50%!
 

Longestday

Helluva Engineer
Featured Member
Messages
2,856
To continue:

Say a team has a bad offense and score 14 points per game in 14 possessions. You play them and they have 9 possession because you eat clock waiting to hike the ball and run the ball keeping the clock moving. They score 7 points. Did the defense do better or worse than their average game? Worse, even though they scored less points per game.

Vise versus...
Say a team has a good offense and scores 42 points per game in 14 possessions because they hurry up and pass. They have 9 possession because they change to a eat clock by wait to hike the ball and run the ball keeping the clock moving. They score 27 points. Did the offense do better or worse than their average game? The same, even though they scored less points per game. 3 points per drive each.
 

Longestday

Helluva Engineer
Featured Member
Messages
2,856
Finally,

If a defense faces a offense that normally scores 50% of the time and only scores 25% of the time for the game, did they do better than other teams? Yes!
If a defense faces a offense that normally scores 25% of the time and scores 50% of the time for the game, did they do better than other teams? No!
 

Vespidae

Helluva Engineer
Messages
5,326
Location
Auburn, AL
"Stats are for losers. The final score is for winners." - Bill Belichick

Belichick's dad wrote the classic text on scouting. Bill is extremely shrewd and you can bet he understands stats. He just doesn't want to tip his hand.

Interesting thread. Teams with a fast offense are often judged harshly from a defensive standpoint because the O is on the field for only a few minutes per game. That's one reason why TOP is no longer a metric with any real importance. It's not uncommon for a D in a fast tempo team to face twice as many plays, so they look "bad" even though if yo adjust the the numbers, they look pretty good.

I think everything is moving to per Drive or per Play for the reasons you mentioned. I recently saw an approach that normalized everything to "per 75 plays" because that was the statistical average of all games played.

Lastly, my observation is that if you have the athletes (like Alabama), you'll grind out a game in the traditional way. Own the LOS, drive 6-7 yards per play, control the clock and let the D do its thing. Run a minimal number of plays.

If you don't have the athletes to own the LOS, you run a fast tempo game - no huddles, short passes, fast, fast fast. Anything to get an edge.

We'll see. Always evolving.
 

Milwaukee

Banned
Messages
7,277
Location
Milwaukee, WI
Belichick's dad wrote the classic text on scouting. Bill is extremely shrewd and you can bet he understands stats. He just doesn't want to tip his hand.

Interesting thread. Teams with a fast offense are often judged harshly from a defensive standpoint because the O is on the field for only a few minutes per game. That's one reason why TOP is no longer a metric with any real importance. It's not uncommon for a D in a fast tempo team to face twice as many plays, so they look "bad" even though if yo adjust the the numbers, they look pretty good.

I think everything is moving to per Drive or per Play for the reasons you mentioned. I recently saw an approach that normalized everything to "per 75 plays" because that was the statistical average of all games played.

Lastly, my observation is that if you have the athletes (like Alabama), you'll grind out a game in the traditional way. Own the LOS, drive 6-7 yards per play, control the clock and let the D do its thing. Run a minimal number of plays.

If you don't have the athletes to own the LOS, you run a fast tempo game - no huddles, short passes, fast, fast fast. Anything to get an edge.

We'll see. Always evolving.

I think we all understand stats.
 

iceeater1969

Helluva Engineer
Messages
9,667
I have believed that our offense efficiency per drive improved as the game progressed. I believed it was not a reversion to an yearly average. Is it a variable that feeds - improves with the number of touches? I think so. Agree?


Early in the Johnson era we seemed to get rolling and win games or come close at the end. Ditto in the 2014 but big time that year. It seemed like the defenses became tired and we made better gains at end of games.
 

bke1984

Helluva Engineer
Messages
3,447
I didn't read the essay because I understand and agree with the concept. But if it wasn't addressed, can someone make a similar post explaining why our defense isn't as good as many think given that they surrender points on just about every drive?
 

ilovetheoption

Helluva Engineer
Messages
2,816
I didn't read the essay because I understand and agree with the concept. But if it wasn't addressed, can someone make a similar post explaining why our defense isn't as good as many think given that they surrender points on just about every drive?

I don't think anybody thinks the defense is any good. FEI (which is essentially a points per drive stat) has GT in the 90's nationally.
 

bke1984

Helluva Engineer
Messages
3,447
Georgia Tech is rated at 34.4 in defensive efficiency. That's 105 out of 128 teams. By comparison, Alabama is number one at 97.9.

There you go...and every week people call into the coaches show and ask about all the "problems" with the offense. I've rarely (if ever) heard someone call in and just go off on why our defense surrenders first downs at will.

It is currently and has been the biggest problem our teams have faced the past 8+ years and we haven't done anything to make me think it's getting better.
 
Top