Missed Tackles

[GaTech4ever]

According to PFF, Tech missed 14 tackles against Louisville. However, three players contributed to > 70% of the missed tackles.

Clayton Powell-Lee: 4 missed tackles (missed 44.4% of tackle attempts)
Trenilyas Tatum: 3 missed tackles (missed 27.3% of tackle attempts)
Kenan Johnson: 3 missed tackles (missed 50% of tackle attempts)

The remaining four missed tackles were one missed tackle each by four different players. This tells me the tackling issue is maybe not as systemic as it seemed during the game, but rather three specific starters who are around the ball a lot and need to tackle better.

 

[GSOJacket]

Not sure I trust PFF (whoever they are). Our secondary looked more like a bunch of matadors than football players against Louisville. We need big improvement here.

 

[57jacket]

According to PFF, Tech missed 14 tackles against Louisville. However, three players contributed to > 70% of the missed tackles.

Clayton Powell-Lee: 4 missed tackles (missed 44.4% of tackle attempts)
Trenilyas Tatum: 3 missed tackles (missed 27.3% of tackle attempts)
Kenan Johnson: 3 missed tackles (missed 50% of tackle attempts)

The remaining four missed tackles were one missed tackle each by four different players. This tells me the tackling issue is maybe not as systemic as it seemed during the game, but rather three specific starters who are around the ball a lot and need to tackle better.

Key added that his team’s defense will continue to focus on tackling this week. The first-year coach said the Jackets have had 400 live reps of scrimmage in the offseason and while he has seen improvement in that area of tackling, it always can stand to be better.

He said the next step for the Jackets’ defense is to tackle with confidence and trust, knowing that there are 10 teammates behind a defender willing and ready to make the play.

“You’ve got to have 10 other people running with their hair on fire ready to have that guy’s back,” Key said. “We’ve worked on that as well.”

 

[LongforDodd]

Not sure I trust PFF (whoever they are). ...

I've always been suspect of their "model". Maybe grading players would be a good application for AI in the future.

 

[slugboy]

I've always been suspect of their "model". Maybe grading players would be a good application for AI in the future.

As someone who has worked with AI commercially, I prefer scouts grading every play. The scouts actually know football.

 

[Gtech50]

As someone who has worked with AI commercially, I prefer scouts grading every play. The scouts actually know football.

PFF graders are not scouts. They use offshore teams to watch film and follow a rubric for grading. Pretty sure most of the people "grading" plays have never played football and probably never even watched it until PFF hired them.

 

[LongforDodd]

As someone who has worked with AI commercially, I prefer scouts grading every play. The scouts actually know football.

I'm not arguing but this PFF has someone watching every play of every player in every game? How many people do they employ to do this?

I'll bet the Chicom AI wizards could come up with something that could eat in to PFF's market.

 

[slugboy]

I'm not arguing but this PFF has someone watching every play of every player in every game? How many people do they employ to do this?

I'll bet the Chicom AI wizards could come up with something that could eat in to PFF's market.

“ Over the last 17 years, PFF has built the world's most comprehensive football database, covering every player on every play of every game for the NFL, FBS and FCS. Its data and tools are trusted by all 32 NFL teams, 131 FBS teams, 47 FCS teams and professional leagues such as the CFL, USFL and XFL.”

Yes, they grade all the players on every play, in all FBS games

There’s a lot more on their services in Wikipedia

Pro Football Focus - Wikipedia

en.wikipedia.org

There’s a solid criticism of their methods here: https://bostonsportsmedia.com/2014/06/04/can-pro-football-focus-stats-be-blindly-trusted/

Keep in mind, that criticism mainly says their analysis isn’t perfect and may have some bias and subjectivity. It’s going to be much better than fan recollections of the game, and valuable enough to most coaches to know where to look.

I can think it’s imperfect and also think it’s one of the best sources of data out there.

As for AI, this is far more true than most people know

 

[mts315]

Key added that his team’s defense will continue to focus on tackling this week. The first-year coach said the Jackets have had 400 live reps of scrimmage in the offseason and while he has seen improvement in that area of tackling, it always can stand to be better.

He said the next step for the Jackets’ defense is to tackle with confidence and trust, knowing that there are 10 teammates behind a defender willing and ready to make the play.

“You’ve got to have 10 other people running with their hair on fire ready to have that guy’s back,” Key said. “We’ve worked on that as well.”

I am not a professional coach but this seems counter intuitive to me. I would think that the mindset of a defender would be, "I better make this tackle because no one is there to help me."

I think the above makes it more likely for a defender to try and make some huge hit and not wrap-up, hoping to get a highlight type play, knowing that they have some back-up if they screw it up.

 

[TromboneJacket]

I am not a professional coach but this seems counter intuitive to me. I would think that the mindset of a defender would be, "I better make this tackle because no one is there to help me."

I think the above makes it more likely for a defender to try and make some huge hit and not wrap-up, hoping to get a highlight type play, knowing that they have some back-up if they screw it up.

I think it depends on the individual and what the thought patterns are. Different people respond differently to the same situation. One guy might respond by thinking “I’m going to make this tackle just like we practiced!” Another might think “Oh man, I better not miss this tackle or else the runner will have a huge gain, and I’ll be letting down the whole team!” The latter is more likely to result in unforced errors. The important thing for success is that the coaches are instilling the proper techniques and keeping the players confident that they can execute them on gameday.

 

[GetYourBuzzOn]

NFL and College Teams pay for PFF's data, so that should say something about their reputation.

 

[Gtech50]

NFL and College Teams pay for PFF's data, so that should say something about their reputation.

Its better than nothing and cheap so thats why teams use it. But it is literally some dudes in India getting paid <$5 per hour to follow a grading rubric. They have no way to know what play was called, did the player actually do his assignment, did the players make the right presnap adjustments, etc.

 

[Root4GT]

As someone who has worked with AI commercially, I prefer scouts grading every play. The scouts actually know football.

PFF is reasonable for Pro games as there are a limited number every week. For college with so many games it is nearly impossible for there to be true consistency in the grading of games and teams. While there may be some value for fans using the info for anything more than online board discussions would be a mistake at the college level.

 

[lv20gt]

I feel like parsing it out by individuals needs a lot more context.

3 starters being inconsistent tacklers is a huge problem. That is 27% of your starters. Even more so because not all positions are going to be getting equal opportunities. For example a space eating DT isn't going to have as many attempts as a LB that he is opening up space for. So that DT having a low number of missed tackles isn't the same as an LB having a missed number of tackles. So it isn't really just 27% in practice as those 3 are likely to account for a higher % of attempts than that.

Those 3 accounted for 26 tackle attempts. According to ESPN as a team we had 39 solo tackles, 18 assisted tackles, and 14 missed tackles. Depending on how you look at it that can mean slightly different things. That is a total of 71 tackling opportunities just adding those together and that would be 19.7% off those being missed. The three players you talked abut account for 36.6% of those opportunities. Now, you could also view it as assisted tackles being multiple players participating in a single opportunity. That would mean 62 opportunities at most. That would put the miss % at 22.5% and those 3 players being a part of 41.9% of plays. This may be slightly high if Powell and Tatum shared any of their assisted tackles. Not a perfect science, as there are a couple things I couldn't quickly find a way to try and account for, but I think it gives a rough picture of the idea. I don't know which way to look at it and either way if you really wanted to see what those numbers mean you'd need to have a that info for the country to see how it stands up.

In any case, I don't think we can put too much stock in it one way or another. We tackled poorly against UL to say the least. If we tackle to that level throughout the year then it will be a big issue no matter how you want to parse it. If it is limited to those three players but it's having as big as an impact as it did against UL, then you can't really write it off as being a limited issue. But, overall, it was one game, and it's possible that it was an outlier. We very well could be a better tackling team than we showed against UL, and it's also possible that we have weaknesses defensively that UL wasn't able to exploit.

 

[takethepoints]

“ Over the last 17 years, PFF has built the world's most comprehensive football database, covering every player on every play of every game for the NFL, FBS and FCS. Its data and tools are trusted by all 32 NFL teams, 131 FBS teams, 47 FCS teams and professional leagues such as the CFL, USFL and XFL.”

Yes, they grade all the players on every play, in all FBS games

There’s a lot more on their services in Wikipedia

Pro Football Focus - Wikipedia

en.wikipedia.org

There’s a solid criticism of their methods here: https://bostonsportsmedia.com/2014/06/04/can-pro-football-focus-stats-be-blindly-trusted/

Keep in mind, that criticism mainly says their analysis isn’t perfect and may have some bias and subjectivity. It’s going to be much better than fan recollections of the game, and valuable enough to most coaches to know where to look.

I can think it’s imperfect and also think it’s one of the best sources of data out there.

As for AI, this is far more true than most people know
View attachment 14706

Since AI has come up, I'll post this:

Three-Toed Sloth

The first post is about Farrell and Shalizi's article calling LLM systems shaggoths. As they point out, LLM systems are just the latest of these: markets, corporations, states, bureaucracies came first and do pretty much the same thing.

I admit this is a bit nerdy, but the idea is sound, imho.

 

[GaTech4ever]

Its better than nothing and cheap so thats why teams use it. But it is literally some dudes in India getting paid <$5 per hour to follow a grading rubric. They have no way to know what play was called, did the player actually do his assignment, did the players make the right presnap adjustments, etc.

I won’t argue with all of this, but in terms of this thread, your comment(a) have nothing to do with being able to count up missed tackles and look at the number of the player who missed them. So it’s kind of irrelevant to this thread. PFF doesn’t miss when it comes to snap counts or easy things like broken tackles, missed tackles, snaps in certain alignments, etc.

 

[57jacket]

Just watch the replay. You can clearly see the missed OPEN FIELD ONE ON ONE tackles. Trying to statistically account for all missed tackles, including more than one tacker is futile.

 

[BuzzStone]

Those stats are much better for us than I would have thought. I know CPL missed at least 7-8 and thats just from memory.

 

[g0lftime]

I feel like parsing it out by individuals needs a lot more context.

3 starters being inconsistent tacklers is a huge problem. That is 27% of your starters. Even more so because not all positions are going to be getting equal opportunities. For example a space eating DT isn't going to have as many attempts as a LB that he is opening up space for. So that DT having a low number of missed tackles isn't the same as an LB having a missed number of tackles. So it isn't really just 27% in practice as those 3 are likely to account for a higher % of attempts than that.

Those 3 accounted for 26 tackle attempts. According to ESPN as a team we had 39 solo tackles, 18 assisted tackles, and 14 missed tackles. Depending on how you look at it that can mean slightly different things. That is a total of 71 tackling opportunities just adding those together and that would be 19.7% off those being missed. The three players you talked abut account for 36.6% of those opportunities. Now, you could also view it as assisted tackles being multiple players participating in a single opportunity. That would mean 62 opportunities at most. That would put the miss % at 22.5% and those 3 players being a part of 41.9% of plays. This may be slightly high if Powell and Tatum shared any of their assisted tackles. Not a perfect science, as there are a couple things I couldn't quickly find a way to try and account for, but I think it gives a rough picture of the idea. I don't know which way to look at it and either way if you really wanted to see what those numbers mean you'd need to have a that info for the country to see how it stands up.

In any case, I don't think we can put too much stock in it one way or another. We tackled poorly against UL to say the least. If we tackle to that level throughout the year then it will be a big issue no matter how you want to parse it. If it is limited to those three players but it's having as big as an impact as it did against UL, then you can't really write it off as being a limited issue. But, overall, it was one game, and it's possible that it was an outlier. We very well could be a better tackling team than we showed against UL, and it's also possible that we have weaknesses defensively that UL wasn't able to exploit.

At some point you have to play guys that will tackle with physicality. There is a tradeoff for pass coverage and speed but DB's are often tackling one on one in space. Have to do both. The really good ones do both.

 

[Gtech50]

I won’t argue with all of this, but in terms of this thread, your comment(a) have nothing to do with being able to count up missed tackles and look at the number of the player who missed them. So it’s kind of irrelevant to this thread. PFF doesn’t miss when it comes to snap counts or easy things like broken tackles, missed tackles, snaps in certain alignments, etc.

I was replying to other comments about PFFs reputability. This thread has gone off topic from the OP.