Postgame: ND 31-GT 13

[Root4GT]

As far as I know, there's only one stat that's 100% aligned with winning--point differential. If you score more than the other guy, you win all the time.

The other stats are there to be predictive and explanatory.

This chart is mostly for fun, and I think it probably got invented because the author kept hearing people say "we were really in this game, but the score doesn't really show it".

So, what he did was take the percentage of plays that each team did something net positive and calculate the difference. It's not scientific, but when you take out acts of god, gusts of wind blowing the ball to the side of the upright, bounces off of someone's helmet for an interception and a pick six, the team that consistently does more things right tends to win. There were 14 games there where the team that consistently won each down still lost the game--because when they did lose the down, they lost it bigtime.

In our case, you could say on offense or defense we won about 42% of the plays, and Notre Dame won about 58% of them--that's a pretty solid win.

That's back of the napkin kinda math, though.

But again, this is mostly "Hey, fan. Yeah, I know that you thought you were in it, but nah, you really weren't".

More scientifically, you can look at post-game win expectancy -- we're at 0.4%. The Notre Dame game wasn't close. Neither were Syracuse or Louisville.

264	19-Oct	Non	0	Georgia Tech	H	Notre Dame	13	31	0	-18	0.4%	-21.3

The games we won, we won solidly. The games we lost, we lost solidly.

====

Now, there's some real stats work here: https://www.footballstudyhall.com/2014/1/24/5337968/college-football-five-factors
The SP+ stats are based on it (and that's explained at https://www.sbnation.com/college-fo...ge-football-advanced-stats-analytics-rankings)

Success rate is what I listed above. I've seen some different definitions, but here are a couple:

Success Rate​

A common Football Outsiders tool used to measure efficiency by determining whether every play of a given game was successful or not. The terms of success in college football: 50 percent of necessary yardage on first down, 70 percent on second down, and 100 percent on third and fourth down.

Success Rate+​

An opponent-adjusted version of Success Rate. As with most other "+" measures, it is built around a baseline of 100.0. Anything over 100.0 is better than average, anything below 100.0 is worse than average.

If you win the Success Rate battle, you win 83% of the time. So, it's a really good predictor of whether you deserved to win (and it doesn't include a lot of the miraculous or "lucky" stuff)
If you win the Explosive Plays battle, you win 86% of the time (but it does include the "lucky" plays that people say "if they hadn't had that pick six, we would have won")
If you win the Finishing the Drive battle, you win 75% of the time (i.e. when you get more points per trip inside the opposing 40)
If you win the Field Position battle (using average starting field position), you win 72 percent of the time.
If you win the Turnover battle (using turnover margin), you win 73 percent of the time.

We didn't win any of those--especially not "Finishing the Drive". The first is being consistent and efficient. The second is getting quick wins. The third is closing the deal.

And, crap, we've been pretty lucky this year, and looking back at it the Syracuse, ND, and Louisville games really show it. As good as our offense is--and it's good--it's not doing as well in success, explosiveness, or finishing as it needs to (it's been good at finishing EXCEPT in three games). Special teams has been losing field position battles for us.

The bad news is, we have a lot of room to improve. The good news is that we can get a LOT better.



Right on, brother.

Good explanation! The truth hurts at times and our 3 losses were games where we were clearly outplayed. You "analysis" matches the eyeball test as well!

 

[bobongo]

The Notre Dame game wasn't close. Neither were Syracuse or Louisville.

The Notre Dame game wasn't close. Anybody could see we were drifting out of it by the middle of the third quarter.

But Syracuse and Louisville were close games well into the fourth quarter as pertains to the scoreboard, which is the only statistic that actually counts. Tech was probably one stop away from beating Syracuse and was right in there against Louisville until the blocked kick with about five minutes to go. I guess we weren't in it according to some metric or other.

 

[cpf2001]

Success rate is fascinating in its popularity currently because it's like the antithesis of the "this isn't built to score quickly" cliche about the Paul Johnson offenses.

Steady-medium-gain success wins a lot of games. That's what that shows. And I think even the best "traditional" offense teams with great success rates are pretty similar to those option teams - they can grind it out against almost any defense, but if the defense screws up, they can break the big play. But if they can't break the big play and have drives stall, or turn it over, or just can't convert in the red zone - they'll find themselves in the "success rate says we should've done much better!!" loss category some.

But I think as a "game report" stat it's very prone to score effects since strategy changes when you're down by two scores. You don't want to wait for the D to break and give up the big play, you're looking for more on every down. And then you lose some of that consistency; meanwhile, slow-and-steady is what your opponent wants, so they're less likely to be pushing. (And even Paul Johnson had a pretty effective two-minute-drill variant of his offense for this - but it wasn't quite as consistent as the base one.)

The Syracuse game especially "suffered" on stats-paper for GT on those. But eventually ended up close because the successful big plays were big enough. The Louisville one too to a bit less of an extent - the scoreboard was closer for longer, so we kept doing what wasn't working, IMO to our detriment, vs changing to a more aggressive offense.

 

[bke1984]

We made too many mistakes on ST, missed tackles, and both OL and DL were dominated against ND. Maybe could have had 6 more points but wouldn't have made any real difference.

We easily could have been +6
They easily could have been -7

That gets us to 19-24. Still lose. But it’s a very close game at that point and you never know what happens.

If we play clean we can be in most games. We need to get back to the way we were at the end of last season.

 

[bke1984]

3 losses were games where we were clearly outplayed.

I agree with Cuse and ND, but not Louisville. I actually felt like we outplayed them in just about every way but a handful of plays where we screwed up royally.

 

[Bogey]

As far as I know, there's only one stat that's 100% aligned with winning--point differential. If you score more than the other guy, you win all the time.

The other stats are there to be predictive and explanatory.

This chart is mostly for fun, and I think it probably got invented because the author kept hearing people say "we were really in this game, but the score doesn't really show it".

So, what he did was take the percentage of plays that each team did something net positive and calculate the difference. It's not scientific, but when you take out acts of god, gusts of wind blowing the ball to the side of the upright, bounces off of someone's helmet for an interception and a pick six, the team that consistently does more things right tends to win. There were 14 games there where the team that consistently won each down still lost the game--because when they did lose the down, they lost it bigtime.

In our case, you could say on offense or defense we won about 42% of the plays, and Notre Dame won about 58% of them--that's a pretty solid win.

That's back of the napkin kinda math, though.

But again, this is mostly "Hey, fan. Yeah, I know that you thought you were in it, but nah, you really weren't".

More scientifically, you can look at post-game win expectancy -- we're at 0.4%. The Notre Dame game wasn't close. Neither were Syracuse or Louisville.

264	19-Oct	Non	0	Georgia Tech	H	Notre Dame	13	31	0	-18	0.4%	-21.3

The games we won, we won solidly. The games we lost, we lost solidly.

====

Now, there's some real stats work here: https://www.footballstudyhall.com/2014/1/24/5337968/college-football-five-factors
The SP+ stats are based on it (and that's explained at https://www.sbnation.com/college-fo...ge-football-advanced-stats-analytics-rankings)

Success rate is what I listed above. I've seen some different definitions, but here are a couple:

Success Rate​

A common Football Outsiders tool used to measure efficiency by determining whether every play of a given game was successful or not. The terms of success in college football: 50 percent of necessary yardage on first down, 70 percent on second down, and 100 percent on third and fourth down.

Success Rate+​

An opponent-adjusted version of Success Rate. As with most other "+" measures, it is built around a baseline of 100.0. Anything over 100.0 is better than average, anything below 100.0 is worse than average.

If you win the Success Rate battle, you win 83% of the time. So, it's a really good predictor of whether you deserved to win (and it doesn't include a lot of the miraculous or "lucky" stuff)
If you win the Explosive Plays battle, you win 86% of the time (but it does include the "lucky" plays that people say "if they hadn't had that pick six, we would have won")
If you win the Finishing the Drive battle, you win 75% of the time (i.e. when you get more points per trip inside the opposing 40)
If you win the Field Position battle (using average starting field position), you win 72 percent of the time.
If you win the Turnover battle (using turnover margin), you win 73 percent of the time.

We didn't win any of those--especially not "Finishing the Drive". The first is being consistent and efficient. The second is getting quick wins. The third is closing the deal.

And, crap, we've been pretty lucky this year, and looking back at it the Syracuse, ND, and Louisville games really show it. As good as our offense is--and it's good--it's not doing as well in success, explosiveness, or finishing as it needs to (it's been good at finishing EXCEPT in three games). Special teams has been losing field position battles for us.

The bad news is, we have a lot of room to improve. The good news is that we can get a LOT better.

Let me ask this. Say we have first down, and are called for holding making it first and 20. So to be successful, on first down, we would need to gain 10 yds and on 2nd down we need 6 yds, making it 3rd and 4, correct?
And say we only gain 5 on first down which would be unsuccessful, we would need to gain 9 yds on our 2nd down again making it 3rd and 4 to be successful?

 

[stinger78]

The Notre Dame game wasn't close. Anybody could see we were drifting out of it by the middle of the third quarter.

But Syracuse and Louisville were close games well into the fourth quarter as pertains to the scoreboard, which is the only statistic that actually counts. Tech was probably one stop away from beating Syracuse and was right in there against Louisville until the blocked kick with about five minutes to go. I guess we weren't in it according to some metric or other.

The ND game was 14-7 when we moved the ball into FG range with 11 secs left and we blew the FG. That would have made the score at the half 14-10.

They scored to make it 21-7 and then 24-7 in the 4th. Game is starting to slide away but not a blowout. We move the ball into scoring position again and botch the second FG. That would have made the score 24-10, a 2-score game. Then we toss the pick-6 with 2 minutes left. 31-7… game. We drive for a score and miss the 2-point conversion. 31-13… could just as easily have been 31-20.

The score looked blowout-ish but the game wasn’t.

 

[slugboy]

Let me ask this. Say we have first down, and are called for holding making it first and 20. So to be successful, on first down, we would need to gain 10 yds and on 2nd down we need 6 yds, making it 3rd and 4, correct?
And say we only gain 5 on first down which would be unsuccessful, we would need to gain 9 yds on our 2nd down again making it 3rd and 4 to be successful?

There are some examples at https://www.actionnetwork.com/ncaaf/success-rate-definition

If you have first and 20, you need to pick up at least 10. If you have second and 12, you need to pick up 9 (8.4) yards.

If you have first and then, and you get five yards on first down, three on second, one on third, and one on fourth, then you have a 50% success rate.

When we would see the flexbone get four on first, four on second, and four on third, that's a 67% clip. I think I've seen it defined by others as 40% of required yards on first down, which is more forgiving. Three yards on first down is definition falling behind the chains.

Edit: man autocorrect is having a field day with my typing lately

 

[slugboy]

There are some examples at https://www.actionnetwork.com/ncaaf/success-rate-definition

If you have first and 20, you need to pick up at least 10. If you have second and 12, you need to pick up 9 (8.4) yards.

If you have first and then, and you get five yards on first down, three on second, one on third, and one on fourth, then you have a 50% success rate.

When we would see the flexbone get four on first, four on second, and four on third, that's a 67% clip. I think I've seen it defined by others as 40% of required yards on first down, which is more forgiving. Three yards on first down is definition falling behind the chains.

Edit: man autocorrect is having a field day with my typing lately

Since we had this conversation



Seven teams doing great at explosive plays and keeping ahead of the chains. Six are in the top 25 and in the CFP discussion. And also Auburn, who is not.

Auburn is miserable at scoring inside the 40 and at turnovers.