What are your measures for success with the new defense?

[Longestday]

How did Roof do it in 2014, but not since?

 

[IEEEWreck]

The real problem with averaging approaches to statistics is that football is a non-homogenous game and the number of games is too small for the law of large numbers.

However defensible the ranking of averages are, the relationship of the average with potential or change needed to improve isn't just nonlinear, it's also not ordinal. That's because unlike in baseball where strategic choices make relatively little difference, football outcomes are massively altered by strategy. In many ways, GTs problem has been being locked into a single strategy on D, and that strategy being relatively ineffective.

Maybe it's true that pure agressively play is better than pure bend but don't break play, but any pure strategy is a liability in a game with diverse opponents and time to adjust in between games.

So, my metric? Some ability to exert control over pace on D. That means something like 3.5 or fewer yd per play in bend style strategy and negative yardage or turnovers 25% of the time in agressive strategy.

 

[Longestday]

Ability to score or prevent a score is, in my opinion, the best averaging stat and predictor of a good offense or defense. Points per game does not work due to the variability of possession numbers (hurry up versus grind it out). The second part of normalization is to compare when playing the same opponents or to adjust based on the competition you play.

If you want to understand why someone is good or bad, then you look into the details. Turnovers, TFL, hurries, passes defended percent, yards per run, and yards per pass, would tell you why you are good or bad on defense.

3rd down is not a great metric as we have seen you could have 70% stops on third down but what if 80% of first of 1st downs don’t require a 3rd down.

 

[g0lftime]

How did Roof do it in 2014, but not since?

Without going back and looking at the stats from that year, it did seem that we got some turnovers that year at key times. VT threw 2 late interceptions when we were behind, GSU lost a pitch that gave us a chance to make a great comeback. MSU had some turnovers in the Orange Bowl but we gave up over 600 yards to them. We got some breaks that year. Was that the year Pitt had all those turnovers? Not sure it was Roof's D. I do think he ran a fundamentally sound defense. It just didn't work well enough with our personnel and he became very predictable with his conservative calls.

 

[BCJacket]

The defense doesn't play in a vacuum. Statistics will be a nice indicator, but the only measure that matters in the history books is W-L.

Bottom line: Does the D do enough to win the game when the offense holds up its end of the bargain? If the D is statistically bad, but make plays when they need to to win the game (eg 2014), I'll be content. Last year the D didn't need to be a whole lot better for us to have won 8+ games, they just needed to make a few plays at the end of halves and OT to stop killer scores.

I'd love eventually for the D to win a game when the offense doesn't hold up its end. Win a game 13-9, with a defensive TD or something. Can't remember the last time that happened...

 

[AE 87]

The defense doesn't play in a vacuum. Statistics will be a nice indicator, but the only measure that matters in the history books is W-L.

Bottom line: Does the D do enough to win the game when the offense holds up its end of the bargain? If the D is statistically bad, but make plays when they need to to win the game (eg 2014), I'll be content. Last year the D didn't need to be a whole lot better for us to have won 8+ games, they just needed to make a few plays at the end of halves and OT to stop killer scores.

I'd love eventually for the D to win a game when the offense doesn't hold up its end. Win a game 13-9, with a defensive TD or something. Can't remember the last time that happened...

I think everybody agrees with this.

Ref to stats is to get an indicator of when the D is holding up it's end.

W-L doesn't tell you that.

 

[slugboy]

Not a stat, but I also care how many times the defense comes off the field either intense or with smiles on their face going "whoo!". I'd like to see them bought in and engaged--even having fun, if that's possible.

 

[bravejason]

How did Roof do it in 2014, but not since?

A change in approach directed by CPJ. The offense was so good that year that it could recover from big plays surrendered by the defense. So CPJ instructed Roof to defend aggressively to increase 3-and-outs and turnovers at the cost of increasing the number of big plays surrendered. I think CPJ said this during a radio show, but possibly it was a press conference.

 

[bravejason]

...
I'd love eventually for the D to win a game when the offense doesn't hold up its end. Win a game 13-9, with a defensive TD or something. Can't remember the last time that happened...

Agree.

GT doesn’t have many such games. Going back to the 12-13 season, I only found three games like that:

BC in ‘16 was 17-14 W
uga in ‘16 was 13-7 L
VT in ‘13 was 17-10 L

In all the other games one team scored at least 20 points.

That said, I think bigger problem is those times when the defense can’t stop the opposing offense. Those occur way too often. Last year that happened a lot and, sadly, happen most years against uga. It’s quite disheartening to know that the opposing team can essentially score at will and that the only way to win is have the last possession and the lead.

 

[RamblinCharger]

1. I don't want to fear 3rd and long
2. FEI on defense needs to improve from 60's or worse to top 45
3. sacks - last year we averaged 1.55 per game. I'd love to average 2.5 a game this year so somewhere around 30-35 sacks as a team would be great improvement

 

[bke1984]

I want us to surrender less points than we score

...in all seriousness though...more turnovers, more sacks (negative plays in general). I'd like to see us in the top-25 in turnover margin and TFL.

I don't have the stats in front of me, but it seems like we've been terrible in the sacks category for quite a while now. Could be in part due to the spread offenses that get rid of the ball quickly, but that leads me to turnovers...if we're rushing the passer and he's dumping the ball fast, then jump the routes, etc. to generate more TO's.

A personal friend of Al Groh's once told me that his philosophy was that if you wait long enough the opposing QB will make a mistake...well unfortunately that's just not true anymore these days with the kids that are identified at age 10 and spend their entire middle and high school years perfecting their passing skills. So IMO there are two ways to play excellent defense these days...1) have the best athletes possible, man up with the DB's and just be better than your opponent, or 2) be super-aggressive and disguise what you do. We fall into bucket #2 and always have. The issue with this style of defense is the variance. Style 1 is consistently good when you have Alabama's squad...style 2 lends itself to blowing up when things aren't going your way.

 

[YJMD]

Ability to score or prevent a score is, in my opinion, the best averaging stat and predictor of a good offense or defense. Points per game does not work due to the variability of possession numbers (hurry up versus grind it out). The second part of normalization is to compare when playing the same opponents or to adjust based on the competition you play.

This makes sense but is incomplete because these things are related to each other. Field position and game situation have significant impact on those measures. I think that impact is magnified for us because we are good at utilizing 4th down conversions past midfield and at/near the top of the NCAA in red zone offense, especially with touchdowns. We possess the ball and limit opportunities to score, making game situation really important to results.

Thus, I care a bit more that the defense creates turnovers and negative plays than improves in absolute points per drive. Obviously, those things are likely to correlate as well.

 

[senoiajacket]

Good stuff.

As you say, per game stats hide our fewer possessions per game. It will make our pts allowed look better than it is and interceptions look worse.

Still awful regardless.

Wow, three year average of 34.3 points per game given up even with fewer possession per game. I would really hate to see that number if the offense we run didn't "make it look better" by holding the ball an so long and eliminating an extra possession or two. Would we have been at 40 pts per game given up if we had a "normal" # of possession per game? Shudder!

 

[lv20gt]

Ultimately the goal is to win so I'm looking for the defense to win. Imo we should be able to score 27 or so against any team so if our defense can be a net 24 points allowed or better we should win. How we get there doesnt really matter. If we give up 35 but get two defensive scores then fine. If we get no turnovers but hold them to 21 then fine.

 

[AE 87]

Ultimately the goal is to win so I'm looking for the defense to win. Imo we should be able to score 27 or so against any team so if our defense can be a net 24 points allowed or better we should win. How we get there doesnt really matter. If we give up 35 but get two defensive scores then fine. If we get no turnovers but hold them to 21 then fine.

In 2017, CU held pwr5 teams to 15.3 pts/game on average; georgie 16.9; VPI 17.5; Duke 22.1; and Miami 22.2.

That's 5 teams on our schedule that are top 30 in pts/game defense vs Pwr5. Making the starting point that our offense should be able to score 27+ on all of them, likely with fewer than average possessions is not reasonable in my opinion.

Of those 5 teams, only georgie was also top 30 in pts/game offense vs Pwr5.

 

[lv20gt]

Wow, three year average of 34.3 points per game given up even with fewer possession per game. I would really hate to see that number if the offense we run didn't "make it look better" by holding the ball an so long and eliminating an extra possession or two. Would we have been at 40 pts per game given up if we had a "normal" # of possession per game? Shudder!

What? We aren't anywhere near 34 on the 3 year average. Were at 25.5 total and 27.7 against p5 teams. Also the fewer possession thing is overblown. It accounts for maybe between one and two extra drives per game. So like a FG. Not a TD.

 

[Jmonty71]

Here are 3 VERY achievable goals.

1. Points per game reduced 6 points
2. Forced punts increased 10%
3. Turnovers for the season up 5

I think all 3 would be a great increase and can be measured.

 

[lv20gt]

In 2017, CU held pwr5 teams to 15.3 pts/game on average; georgie 16.9; VPI 17.5; Duke 22.1; and Miami 22.2.

That's 5 teams on our schedule that are top 30 in pts/game defense vs Pwr5. Making the starting point that our offense should be able to score 27+ on all of them, likely with fewer than average possessions is not reasonable in my opinion.

Of those 5 teams, only georgie was also top 30 in pts/game offense vs Pwr5.

And I think with how much our program invests into that one side of the ball it isn't unreasonable to expect that.

 

[AE 87]

And I think with how much our program invests into that one side of the ball it isn't unreasonable to expect that.

What does "with how much our program invests into that one side of the ball" mean?

 

[Longestday]

I think he means with our "faith" in CPJ and how much time CPJ puts into that side of the ball versus a head coach and a O and D coordinator.

Improved D helps field positions and more possessions for more positive turnovers... This helps the O.