Duke GT Notes

[Longestday]

Duke has played all 4 of out of conference teams and one ACC team with a 4 and 1 record
GT has played 3 out of conference teams and 2 ACC teams with a 5 and 0 record

Dukes teams have win loss record of 7 and 19
GT teams have a win loss record of 15 and 13

Duke is ranked 31st with 36.8 points for and 8th with 13.6 points against
GT is ranked 41st with 34.6 points for and 58th with 23.8 points against

GT had 8 drives with Miami
Duke had 16 drives with Miami

Given GT's limited drives, the points for underestimates our team and points against over estimates our team compared to Duke. Dukes strength of schedule over estimates compared to GT.

Duke's turn over margin is 4 and GT's is 5

Duke has 3 interceptions off their QB
GT has 1 interception off their QB

GT 38-21 Tulane
Duke 47-13 Tulane

GT 28-17 Miami
Duke 10-22 Miami

 

[mts315]

GT 38-21 Tulane
Duke 47-13 Tulane

GT 28-17 Miami
Duke 10-22 Miami

Should note that GT played Tulane at Tulane and Duke played them at Duke. Also, that GT played Miami at Home and Duke played them at Miami. Not sure it makes a large difference but home field advantage should factor in.

 

[AE 87]

The typical measure of Scoring Offense (Points/Game) is a raw stat that doesn't account for opposition strength. As is the better Points/Drive stat (which I calculate with a some minor potential for error).

In PPD offense, GT is #11 at 3.31 ppd, and Duke is #38 at 2.63 ppd.
In PPD defense, GT is #92 at 2.43 ppd and Duke is #8 at 1.06 ppd.

 

[dressedcheeseside]

There's no crying in baseball and there's no transitive property in football.

 

[JDjacket]

I knew Duke's schedule was bad especially seeing Troy on the schedule, but Duke's only game against a team with a winning record was Miami and they lost. Tulane is the only team we've played who current has a losing record. Also interesting that Duke's defense only seems to have held most teams to their avg. points scored per season rather than below their avg with the exception of Miami.

There may not be a transitive property in football, but I hope playing against better competition has definitely prepared us much better for this coming game than playing against worse competition like Duke has.

 

[vamosjackets]

Looks to me like Duke's defense has done noticeably better than ours so far. This would be especially noticeable in the Miami and Tulane games. Both teams would've probably had significantly more possessions against Duke than against us. And, as AE87 points out, there's the statistical evidence put forward in a more digestable fashion. We're pretty bad in ppd and they're pretty good. There are some outside factors there like our fumble and score to open the Tulane game, and perhaps Miami's desire to control the clock with a lead against Duke vs a desire to score as quickly as possible against us being behind. I'd lean towards our offense being a bit better than theirs and their defense being a bit better than ours. Their off-week will help them a lot. Our home-field should help us some.

So, ... I'm still saying we need to bring it at pretty much the same level as against UM and VT in order to win this game.

 

[Bruce Wayne]

Looks to me like Duke's defense has done noticeably better than ours so far. This would be especially noticeable in the Miami and Tulane games. Both teams would've probably had significantly more possessions against Duke than against us. And, as AE87 points out, there's the statistical evidence put forward in a more digestable fashion. We're pretty bad in ppd and they're pretty good. There are some outside factors there like our fumble and score to open the Tulane game, and perhaps Miami's desire to control the clock with a lead against Duke vs a desire to score as quickly as possible against us being behind. I'd lean towards our offense being a bit better than theirs and their defense being a bit better than ours. Their off-week will help them a lot. Our home-field should help us some.

So, ... I'm still saying we need to bring it at pretty much the same level as against UM and VT in order to win this game.

Absolutely agreed! the coach speak of "you ain't done nothing yet" and "each game just gets better" is more true than usual this year. This team has to continue to practice with a mission and play with concentration and maximum effort to win every game on the schedule.

They could eventually become a fairly dominant team (like next year or two) but they are nowhere near that point yet, and frankly I don't believe a Tech team could ever achieve that status any other way than having a whole team that tries to get better every single day and maximize their individual potential. There are just not enough athletic 5 star "freaks" of nature types nor are our players at a program that lets you completely ignore being a student so you can spend all your time on a practice field or in a video room.

 

[AlabamaBuzz]

Their D thus far is giving up 4.7 yds per play and ours is giving up 6.4 yds per play...hopefully, we give them very few offensive plays.....

 

[orientalnc]

Data are good, but they ain't perfect. Our yards per play averages are skewed by a few plays early in the season. If you remove the one 92 yard TD in the Wofford game, the one 61 yard TD at Tulane, then the three second half breakdowns against GSU, our yards per play average drops a whole yard per play. The adjustments Coach Roof made after the GSU game were significant.

Another interesting stat is the number of plays our opponents have run. Wofford, Tulane, and GSU all ran 60 plays against us. Then VT ran 72 plays, while Miami only had 44. In reviewing these games, the major difference is our offense. VT got us off the field a lot quicker than Miami. In the first three games we dominated the opponent defensively until we gave up the huge plays. Those huge plays count, so do not miss my point. I think we suffered from missed assignments early because of so many new players on defense, especially the DBs. They have learned from those mistakes and the adjustments Roof made certainly helped in that. Also, have the Austins played since GSU?

I am amazed we beat Miami. They seem to have better athletes at almost every position, but do not play together as a team nearly as well. Al Golden is in trouble down there and could be history if things don't change soon.

 

[GTNavyNuke]

The typical measure of Scoring Offense (Points/Game) is a raw stat that doesn't account for opposition strength. As is the better Points/Drive stat (which I calculate with a some minor potential for error).

In PPD offense, GT is #11 at 3.31 ppd, and Duke is #38 at 2.63 ppd.
In PPD defense, GT is #92 at 2.43 ppd and Duke is #8 at 1.06 ppd.

AE, where did you get the ppd data or did you calculate it?

Football Outsiders will start adjusting for opponent strength after week 7 I believe. Until then, we have less informative data than we will. It's hard to compare the raw numbers as Longest points out since the quality of our opponents has been higher than Dukes. How much? Well, you need the weak transitive property for that.

For DCS, there is a weak transitive property in football (I know you know that). But due to the weakness, you have to compare all teams that play against each other and minimize the error ...... look at the probability of win chart during our game against Miami at http://www.fromtherumbleseat.com/2014/10/6/6916199/advanced-box-score-miami . Ain't no way you can predict the outcome very accurately in advance of the game, it's hard even during the game!

 

[GTNavyNuke]

Here's something that quantifies Strength of Schedule to date. Looking at YTD (which is SOS to date), GT's opponents have been rated .404 and Duke's .185. (.404 means the opponents have a 40.4% chance of beating the average team on a neutral field). I personally think that GSU was / is underrated and will climb during the season.
http://www.jhowell.net/cf/cf2014.htm

We are ranked 21st (.718) and Duke is 46th overall (.607). For GT, .718 is pretty high historically .......

What I don't like is that we are 6-0 versus Duke under CPJ. Eventually Duke will win one ..... probably because they recruit superior talent since players don't like CPJ's scheme (tic).

 

[DTGT]

What I don't like is that we are 6-0 versus Duke under CPJ. Eventually Duke will win one ..... probably because they recruit superior talent since players don't like CPJ's scheme (tic).

Past history does not guarantee future performance, but it is a d*** good guide. Until I see a game where they can hang with us, I will continue believing that they can't beat us when our players are healthy and full of piss and vinegar.

 

[00Burdell]

What I don't like is that we are 6-0 versus Duke under CPJ. Eventually Duke will win one.

The sun will never rise in the west.

 

[AE 87]

AE, where did you get the ppd data or did you calculate it?

Football Outsiders will start adjusting for opponent strength after week 7 I believe. Until then, we have less informative data than we will. It's hard to compare the raw numbers as Longest points out since the quality of our opponents has been higher than Dukes. How much? Well, you need the weak transitive property for that.

For DCS, there is a weak transitive property in football (I know you know that). But due to the weakness, you have to compare all teams that play against each other and minimize the error ...... look at the probability of win chart during our game against Miami at http://www.fromtherumbleseat.com/2014/10/6/6916199/advanced-box-score-miami . Ain't no way you can predict the outcome very accurately in advance of the game, it's hard even during the game!

I calculate it.
Points = 7*sum(rush td, pass td) + 3*sum(FGs)
Drives=sum(rush td, pass td, fg att, punts, lost turnovers, failed 4th dn)

 

[96Jacket]

I calculate it.
Points = 7*sum(rush td, pass td) + 3*sum(FGs)
Drives=sum(rush td, pass td, fg att, punts, lost turnovers, failed 4th dn)

Probably splitting hairs here but wouldn't you include drives that get cut short with the end of a half? In some cases, like the ending of the Miami game when we were in the victory formation, it would skew the numbers in a misleading fashion but in other cases it could affect the ppd stat in a relevant way.

 

[DTGT]

Probably splitting hairs here but wouldn't you include drives that get cut short with the end of a half? In some cases, like the ending of the Miami game when we were in the victory formation, it would skew the numbers in a misleading fashion but in other cases it could affect the ppd stat in a relevant way.

Only if as time expired you got a TD, FGA, punt, TO, or failed 4th down.

 

[AE 87]

Probably splitting hairs here but wouldn't you include drives that get cut short with the end of a half? In some cases, like the ending of the Miami game when we were in the victory formation, it would skew the numbers in a misleading fashion but in other cases it could affect the ppd stat in a relevant way.

At a certain point, it becomes too complicated for a simple stat. Any simple raw measure will have problems. Tifwiw

 

[00Burdell]

Probably splitting hairs here but wouldn't you include drives that get cut short with the end of a half? In some cases, like the ending of the Miami game when we were in the victory formation, it would skew the numbers in a misleading fashion but in other cases it could affect the ppd stat in a relevant way.

Interesting question but, for example, if a possession is ended by the halftime whistle (as opposed to one of the other possession terminating events) then its not really a possession that has bearing on the stat. If the other guy punts to us with three seconds left in the half then that possession can't reasonable be expected to yield points but, if included, would water down points per possession. In other words, I'm more comfortable with the stat if it omits possessions that are either so short they can't terminate naturally or that we had no incentive to try to score during.

 

[ibeattetris]

I guess you don't have access to 2 pt conversion attempt and success? It seems like for total points you'd want to do 6*(totalTd) + 2*twoPtSuccess + epSuccess

Not sure if you have access to the two point or extra point data but a team like Oregon who prefers to go for two would potentially have a different ppd assuming all td's are 7.

Edit: math

 

[collegeballfan]

Here's something that quantifies Strength of Schedule to date. Looking at YTD (which is SOS to date), GT's opponents have been rated .404 and Duke's .185. (.404 means the opponents have a 40.4% chance of beating the average team on a neutral field). I personally think that GSU was / is underrated and will climb during the season.
http://www.jhowell.net/cf/cf2014.htm

We are ranked 21st (.718) and Duke is 46th overall (.607). For GT, .718 is pretty high historically .......

What I don't like is that we are 6-0 versus Duke under CPJ. Eventually Duke will win one ..... probably because they recruit superior talent since players don't like CPJ's scheme (tic).

As Coach says - we will get Duke's best shot. Here from the Durham paper yesterday:

"Ending the Georgia Tech streak – Duke hasn’t defeated Georgia Tech since 2003, making the Yellow Jackets the only ACC Coastal Division team that the Blue Devils haven’t beaten since Cutcliffe came to Durham for the 2008 season.

Cutcliffe admitted ending the streak does carry added importance to him.

“Yes it does, I’m not going to lie to you guys,” Cutcliffe said. “It’s no fun to be 0-6 against anybody. It’s not a habit you want to start. But if you don’t like it you better do something about it. That certainly bothers me and I think it would anybody.”