PPD and our Remaining Games

AE 87

Helluva Engineer
Messages
13,015
Okay, so again the PPD used in this post is a calculated stat. Points = 7x(Rush TD's + Pass TD's) + 3xFG's made. Drives = Rush TD's + Pass TD's + FG's Att + Punts + Failed 4th + lost turnovers. I calculate an OPPD for our offense and a DPPD for our points allowed. As a rule of thumb, teams playing more than 2 Pwr 5 opponents average 2.15 pts/drive and allow 2.32 points/drive in those games.

Here are our OPPD for each of our ACC games compared to the opponent's average DPPD vs Pwr5 opponents:

upload_2018-11-5_15-11-48.png


There's a pretty good indicator of Bad GT vs Good GT on offense. Against Pitt and Duke, we scored 2/3 and less of the points per drive they allow on average. Against CU and LOLvl about 70% more efficiently than they typically allow. VPI game was crazy good and the unc game was alright but turnovers slowed us down.

If we allow that the VPI game was an aberration and that the unc game was marked by unusual fumbles, I think 1.7 above opposition averages seems a reasonable assessment.

Now here are our DPPD for each of these games compared to our opponent's average OPPD vs Pwr5 opponents:

upload_2018-11-5_15-16-20.png


Besides the Pitt game, we've been consistently below average. Now, if we allow that the LOLvl and VPI D may have given up some points because of GT's lead, then I think that saying we allow 1.2 times more points/drive than average is a reasonable assessment.

Now here are our remaining opponents and their average points/drive vs power5 opponents:
upload_2018-11-5_15-28-29.png


Using the 1.7 and 1.2 multipliers, and assuming 12 drives, the predicted scores of these games are:
GT 32 - d'ohU 22
GT 37 - uva 29
GT 37 - u[sic]ga 44

Now, obviously the stat is descriptive and is not intended to be predictive. For example, it doesn't consider run-back scores at all. Still, I thought it was interesting.
 

AE 87

Helluva Engineer
Messages
13,015
I'll take that for the next 2 games please and then we'll re-evaluate. :D

One of the things I find interesting is that Pitt has one of the worst D's and Duke one of the better D's that we've faced based on this stat. Yet, those were the Bad GT games with nearly identical ratio.

CU has the best D and LOLvl the absolute worst D we've faced, and yet our Offensive performance as a ratio against their average was nearly identical.

So, based on this stat anyway, the appearance of Bad GT or Good GT seems independent of opposition. Maybe there are other stats that would show differently.
 

Deleted member 2897

Guest
So, based on this stat anyway, the appearance of Bad GT or Good GT seems independent of opposition.

That's exactly the rub and what's frustrating about this team. We have about the same winning percentage against Florida State, Clemson, and Virginia Tech over the last 10 years that we do with Duke, UNCheat, Virginia, and Pittsburgh. Figure that out.
 

Longestday

Helluva Engineer
Featured Member
Messages
2,856
I think Pitt and Duke PPD may have been effected by injury to TQM. Just like JT playing when Byerly should have came in earlier for Duke 2014.
 

Deleted member 2897

Guest
I would like to point out that Clemson's offense got significantly better once Trevor Lawrence started. Clemson has averaged over 50pts per game since the GT game.

If he doesn’t get injured, he will be one of the best college quarterbacks ever. His football IQ to me is similar to Peyton Manning, but his mobility and throw inventory are even better.
 

iceeater1969

Helluva Engineer
Messages
8,916
Okay, so again the PPD used in this post is a calculated stat. Points = 7x(Rush TD's + Pass TD's) + 3xFG's made. Drives = Rush TD's + Pass TD's + FG's Att + Punts + Failed 4th + lost turnovers. I calculate an OPPD for our offense and a DPPD for our points allowed. As a rule of thumb, teams playing more than 2 Pwr 5 opponents average 2.15 pts/drive and allow 2.32 points/drive in those games.

Here are our OPPD for each of our ACC games compared to the opponent's average DPPD vs Pwr5 opponents:

View attachment 4430

There's a pretty good indicator of Bad GT vs Good GT on offense. Against Pitt and Duke, we scored 2/3 and less of the points per drive they allow on average. Against CU and LOLvl about 70% more efficiently than they typically allow. VPI game was crazy good and the unc game was alright but turnovers slowed us down.

If we allow that the VPI game was an aberration and that the unc game was marked by unusual fumbles, I think 1.7 above opposition averages seems a reasonable assessment.

Now here are our DPPD for each of these games compared to our opponent's average OPPD vs Pwr5 opponents:

View attachment 4431

Besides the Pitt game, we've been consistently below average. Now, if we allow that the LOLvl and VPI D may have given up some points because of GT's lead, then I think that saying we allow 1.2 times more points/drive than average is a reasonable assessment.

Now here are our remaining opponents and their average points/drive vs power5 opponents:
View attachment 4432

Using the 1.7 and 1.2 multipliers, and assuming 12 drives, the predicted scores of these games are:
GT 32 - d'ohU 22
GT 37 - uva 29
GT 37 - u[sic]ga 44

Now, obviously the stat is descriptive and is not intended to be predictive. For example, it doesn't consider run-back scores at all. Still, I thought it was interesting.
Wow! This is a detailed study that has measurable and trackable output that can be recorded over time.
 

AE 87

Helluva Engineer
Messages
13,015
@AE 87 I wonder what those DPPD and OPPD would look like without GT data...

Yeah, I thought that. It would be better math for the projection to future games.

Typically, I don't have the data available since I don't use per game data for calculating ppdvp5 for all 130 teams.

However, I have GT's per game data now so I'll look at it when I have some time.
 
Top