New "QB support" stat being developed

For all things Chicago Bears

Moderator: wab

User avatar
wab
Mod
Posts: 29414
Joined: Sun Nov 02, 2008 12:49 pm
Has thanked: 112 times
Been thanked: 1779 times

This isn't Bears specific, but with all the talk about Fields and how much help he does/doesn't get, I wonder what this stat will show. It's an interesting metric that they are creating.

User avatar
The Cooler King
Head Coach
Posts: 4907
Joined: Wed Dec 18, 2019 11:07 pm
Has thanked: 1114 times
Been thanked: 320 times

Pretty interesting. Lots of data to try and absorb (and possibly pick apart, I'm sure).
User avatar
dplank
Hall of Famer
Posts: 11676
Joined: Tue Nov 29, 2016 9:19 am
Has thanked: 1122 times
Been thanked: 1948 times

Interesting! At a glance, looking at the quadrant chart and pausing it there for a moment, what jumps out at me most is how the top right and bottom left quadrants have 90% of the players. Meaning, most of the efficient QB's have strong supporting casts, and most of the inefficient QB's have poor supporting casts. There were very few QB's that outperformed their surrounding casts, and there were very few QB's that underperformed their supporting cast.

This is the main takeaway, and aligns well with common sense while simultaneously highlighting how important supporting cast is overall. Football is still a team sport, and as important as QB is it's not everything.

Now I'll sit back and wait for folks to spin the data towards individual preferences and biases.
User avatar
Bears Whiskey Nut
Hall of Famer
Posts: 10913
Joined: Tue Aug 21, 2012 4:06 am
Location: Oak Park, IL
Has thanked: 76 times
Been thanked: 462 times

It's a fascinating take. There are going to be some very subjective numbers in there, but I guess there are with all stats like this. For example will they take into account OL penalties, such as holding and false starts as a metric to show how many yards the QB lost in a game? Such as plays were the QB hit a receiver on 50 yard bomb for a TD, only to be called back by a holding penalty. Is that what they are analyzing, or is it something more nebulous than that?
Image
User avatar
Moriarty
Hall of Famer
Posts: 6669
Joined: Sun Aug 28, 2016 1:22 pm
Has thanked: 365 times
Been thanked: 643 times

dplank wrote: Tue Feb 06, 2024 12:15 pm what jumps out at me most is how the top right and bottom left quadrants have 90% of the players. Meaning, most of the efficient QB's have strong supporting casts, and most of the inefficient QB's have poor supporting casts. There were very few QB's that outperformed their surrounding casts, and there were very few QB's that underperformed their supporting cast.
That's one possible explanation.

The other one is - the metric is generally unsuccessful at separating the QB influence from the Everybody Else influence. It tends to just rate "everyone good" or "everyone bad", because it's unable to distinguish, as it was hoped to be able to.
1999-2002: Mouth Off Sports Forum (RIP)
2002-2014: KFFL (RIP)
2014-2016: USAToday Fantasy Sports Forum (RIP)

Hello, my name is Moriarty. I have come to kill your website, prepare to die.
Magilla_Gorilla
Journeyman
Posts: 156
Joined: Thu Apr 02, 2009 3:54 pm
Has thanked: 63 times
Been thanked: 86 times

Here are the two charts shown in the video:

Image
Image
User avatar
The Cooler King
Head Coach
Posts: 4907
Joined: Wed Dec 18, 2019 11:07 pm
Has thanked: 1114 times
Been thanked: 320 times

dplank wrote: Tue Feb 06, 2024 12:15 pm Interesting! At a glance, looking at the quadrant chart and pausing it there for a moment, what jumps out at me most is how the top right and bottom left quadrants have 90% of the players. Meaning, most of the efficient QB's have strong supporting casts, and most of the inefficient QB's have poor supporting casts. There were very few QB's that outperformed their surrounding casts, and there were very few QB's that underperformed their supporting cast.

This is the main takeaway, and aligns well with common sense while simultaneously highlighting how important supporting cast is overall. Football is still a team sport, and as important as QB is it's not everything.

Now I'll sit back and wait for folks to spin the data towards individual preferences and biases.
Yes it's largely what you'd expect but also I wonder if it's struggling to capture the QBs contributions to some of these stats even though it attempts to design around that.

Like outside of Kelce, no one thought highly of Mahomes weapons going into this year. Yet the perfect coverage rank for them is top 10. Were we way off on evaluating the talent he had at WR or does perfectly covered rate implicitly fail to fully separate the QBs impact? I assume it does, but to which degree could have big impact on the meaningfulness of a stat like this. You'd almost have to drill down to a player level contribution score to isolate and test the predictive value of that.
User avatar
dplank
Hall of Famer
Posts: 11676
Joined: Tue Nov 29, 2016 9:19 am
Has thanked: 1122 times
Been thanked: 1948 times

The Cooler King wrote: Tue Feb 06, 2024 12:41 pm
dplank wrote: Tue Feb 06, 2024 12:15 pm Interesting! At a glance, looking at the quadrant chart and pausing it there for a moment, what jumps out at me most is how the top right and bottom left quadrants have 90% of the players. Meaning, most of the efficient QB's have strong supporting casts, and most of the inefficient QB's have poor supporting casts. There were very few QB's that outperformed their surrounding casts, and there were very few QB's that underperformed their supporting cast.

This is the main takeaway, and aligns well with common sense while simultaneously highlighting how important supporting cast is overall. Football is still a team sport, and as important as QB is it's not everything.

Now I'll sit back and wait for folks to spin the data towards individual preferences and biases.
Yes it's largely what you'd expect but also I wonder if it's struggling to capture the QBs contributions to some of these stats even though it attempts to design around that.

Like outside of Kelce, no one thought highly of Mahomes weapons going into this year. Yet the perfect coverage rank for them is top 10. Were we way off on evaluating the talent he had at WR or does perfectly covered rate implicitly fail to fully separate the QBs impact? I assume it does, but to which degree could have big impact on the meaningfulness of a stat like this. You'd almost have to drill down to a player level contribution score to isolate and test the predictive value of that.
Yea really hard to say, no matter how hard they try to make this a stat based POV, there's always subjectivity in how the underlying data is derived. I like the concept, but I doubt it's any more meaningful than anything else ultimately.

I find the Mahomes example interesting, because for all the angst about how bad his WR corps was - how much did it really change from the year before when he put up ungodly numbers? Very little really. JuJu out and Rice in? JMO, but the main delta in KC this year was Bienemy leaving and Nagy, presumably, having more of a hand in the offense.
User avatar
The Marshall Plan
Hall of Famer
Posts: 8187
Joined: Sun Mar 09, 2014 11:47 am
Location: Parts Unknown
Has thanked: 835 times
Been thanked: 1151 times

Oh good.

I can’t wait until these are taken out of context, used to push predetermined narratives and causes untold hours of arguing.
I Take The Field With Fields

Image
User avatar
Bears Whiskey Nut
Hall of Famer
Posts: 10913
Joined: Tue Aug 21, 2012 4:06 am
Location: Oak Park, IL
Has thanked: 76 times
Been thanked: 462 times

The Marshall Plan wrote: Tue Feb 06, 2024 1:44 pm Oh good.

I can’t wait until these are taken out of context, used to push predetermined narratives and causes untold hours of arguing.
Ummmm...that's what we do here dude.
Image
User avatar
BreadNCircuses
Assistant Coach
Posts: 504
Joined: Tue Sep 10, 2019 1:34 pm
Has thanked: 26 times
Been thanked: 82 times

The Marshall Plan wrote: Tue Feb 06, 2024 1:44 pm Oh good.

I can’t wait until these are taken out of context, used to push predetermined narratives and causes untold hours of arguing.
Just think -- you can get in on the ground floor here, and we could potentially quantify the hours arguing from the beginning -- they don't HAVE to remain untold!

:)
2023 Preseason Downside prediction:
5-6 wins, never really healthy all season, a constant shuffling.
We're potentially in a position to draft in the Top 5 again, depending on the Carolina team, and probably have a low-teens (or better) pick ourselves.
User avatar
The Marshall Plan
Hall of Famer
Posts: 8187
Joined: Sun Mar 09, 2014 11:47 am
Location: Parts Unknown
Has thanked: 835 times
Been thanked: 1151 times

BreadNCircuses wrote: Tue Feb 06, 2024 2:07 pm
The Marshall Plan wrote: Tue Feb 06, 2024 1:44 pm Oh good.

I can’t wait until these are taken out of context, used to push predetermined narratives and causes untold hours of arguing.
Just think -- you can get in on the ground floor here, and we could potentially quantify the hours arguing from the beginning -- they don't HAVE to remain untold!

:)
Then I can make a secondary stat tracking the situations we use the secondary stats in!
I Take The Field With Fields

Image
User avatar
Arkansasbear
Head Coach
Posts: 4633
Joined: Tue Jun 16, 2020 10:41 am
Has thanked: 415 times
Been thanked: 598 times

What did Mac Jones have that gave him "lots of help?"

That team is devoid of talent.
User avatar
BreadNCircuses
Assistant Coach
Posts: 504
Joined: Tue Sep 10, 2019 1:34 pm
Has thanked: 26 times
Been thanked: 82 times

The Marshall Plan wrote: Tue Feb 06, 2024 2:12 pm
BreadNCircuses wrote: Tue Feb 06, 2024 2:07 pm

Just think -- you can get in on the ground floor here, and we could potentially quantify the hours arguing from the beginning -- they don't HAVE to remain untold!

:)
Then I can make a secondary stat tracking the situations we use the secondary stats in!
STATCEPTION

Image
2023 Preseason Downside prediction:
5-6 wins, never really healthy all season, a constant shuffling.
We're potentially in a position to draft in the Top 5 again, depending on the Carolina team, and probably have a low-teens (or better) pick ourselves.
User avatar
southdakbearfan
Head Coach
Posts: 4482
Joined: Wed Jul 20, 2011 11:23 pm
Location: South Dakota
Has thanked: 623 times
Been thanked: 285 times

Looking at where a bunch of QBs, like Mac Jones landed, it would seem that it may need a tweak or two.
User avatar
docc
Head Coach
Posts: 3772
Joined: Sun Sep 11, 2011 4:33 pm
Location: Outpost of Reality S.E. Arizona
Has thanked: 860 times
Been thanked: 151 times

How much does Carson Beck make in NIL money?
Beck is one of the most notable figures in college sports, at least in terms of net worth. According to On3, Beck has an NIL valuation of over $1.1 million.

That number is only expected to rise after he finished the 2023 season with a 72.4-percent completion rate, 3,941 passing yards, 24 touchdowns, and six interceptions while leading Georgia to a 13-1 record and Orange Bowl win.
Beck joins a long list of Georgia quarterbacks in both statical success and value as he looks to guide the Bulldogs to another national title. Other notable Georgia quarterbacks include JT Daniels, who once held a net worth of $7 million according to multiple outlets, currently listed at $3 million.

Former walk-on and current Rams quarterback Stetson Bennett's net worth also sits at $3 million. Based on his current trajectory, Beck could surpass both of them by a decent margin as he looks to carry on the prestigious legacy.
------------------------------------------------------------------------------------------------------------------------------------------------------------
Carson Beck A name to remember for next draft. I've seen several of his games. Even the small size will lead to better estimation of more upside 2025.
User avatar
crueltyabc
Hall of Famer
Posts: 5065
Joined: Mon May 04, 2009 7:36 pm
Location: Dallas TX
Has thanked: 74 times
Been thanked: 210 times

If they were using PFF grades @southdakbearfan and @Arkansasbear to determine Mac's support, then the answer is that they like the Oline and they like Demario Douglas

Oline
Trent Brown 80.2
Cole Strange 64.6
David Andrews 71.2
Sidy Sow 64.4
Mike Onwenu 71.5

Weapons
Demario Douglas 74.4
Hunter Henry 69.3
Zeke Elliot 67.6
xyt in the discord chats
User avatar
Arkansasbear
Head Coach
Posts: 4633
Joined: Tue Jun 16, 2020 10:41 am
Has thanked: 415 times
Been thanked: 598 times

crueltyabc wrote: Wed Feb 07, 2024 2:59 pm If they were using PFF grades @southdakbearfan and @Arkansasbear to determine Mac's support, then the answer is that they like the Oline and they like Demario Douglas

Oline
Trent Brown 80.2
Cole Strange 64.6
David Andrews 71.2
Sidy Sow 64.4
Mike Onwenu 71.5

Weapons
Demario Douglas 74.4
Hunter Henry 69.3
Zeke Elliot 67.6
Well I think PFF stats pretty much suck so that could explain it all.

Granted I think coming up with a way to truly grade OL play is next to impossible.
User avatar
dplank
Hall of Famer
Posts: 11676
Joined: Tue Nov 29, 2016 9:19 am
Has thanked: 1122 times
Been thanked: 1948 times

Arkansasbear wrote: Wed Feb 07, 2024 3:15 pm
crueltyabc wrote: Wed Feb 07, 2024 2:59 pm If they were using PFF grades @southdakbearfan and @Arkansasbear to determine Mac's support, then the answer is that they like the Oline and they like Demario Douglas

Oline
Trent Brown 80.2
Cole Strange 64.6
David Andrews 71.2
Sidy Sow 64.4
Mike Onwenu 71.5

Weapons
Demario Douglas 74.4
Hunter Henry 69.3
Zeke Elliot 67.6
Well I think PFF stats pretty much suck so that could explain it all.

Granted I think coming up with a way to truly grade OL play is next to impossible.
It is. My method seems to be as good as any other lol - watch the game with your eye trained on the LOS (vs the ball) and see who is pushing who around. Super scientific! I swear you will see that the team that wins this Uber simple eye test wins the game like 80%+ of the time.
User avatar
Ditka’s dictaphone
Head Coach
Posts: 3841
Joined: Sun Apr 04, 2021 12:33 pm
Has thanked: 668 times
Been thanked: 800 times

Just what the world needs, people to “invent” new stats :frustrated:

:D
(26/09/2023) Winner of the inaugural

Image
User avatar
malk
Head Coach
Posts: 3589
Joined: Wed Aug 22, 2012 7:10 am
Has thanked: 128 times
Been thanked: 182 times

From an initial look at this I think the composite scores they come to are just too far abstracted from actuality to be meaningful. PFF pass block grades aren't great, the perfectly covered rate seems a bit iffy, how defence feeds into offensive performance doesn't appear to be adequately controlled for...

So that combined with how the results look pretty meh, my gut tells me it needs to be dug into more to find out where a structural flaw is.

Just thinking aloud, use of EPA for QBs makes sense but the chart that is included doesn't show where receivers come into play? It must be factored somewhere but it mentions rushing, pass blocking and scheme + WRs and is evenly weighted between those 3 (but there are 4 lol).

Good idea in general but would reserve judgement until the black box is opened up a bit more.
"I wouldn't take him for a conditional 7th. His next contract will pay him more than he could possibly contribute.".

Noted Brain Genius Malk, Summer 2018.

(2020 update, wait, was I right...)
User avatar
mmmc_35
Hall of Famer
Posts: 6092
Joined: Sun Aug 26, 2012 12:25 am
Has thanked: 99 times
Been thanked: 83 times

I really like the idea. The problem with any stat is it is like your brother in law it will lie when he has to.

Everything stat has to be taken as a gauge with your eyes evaluation.

PFF stats. People hate on PFF. I think when you see a player, you gauge his play but its subjective. PFF can bring you closer to reality. PFF is also subjective.

Let's take gervon Dexter. My eyes tell me hes a force. He progressed. I believe he played well. PFF says he played poorly. We can make excuses for the grade. I saw a young player really developing. He played very well at times. I give him a B grade. PFF grades him like a D.
last year he probably therefore a C player. With upside.

I think we all viewed the overall offense with some holes. Mooney who I like didnt play great. The center play was poor. Play calling poor. Individually I think they have some pieces. I think this stat kind of verifies that sentiment.
User avatar
malk
Head Coach
Posts: 3589
Joined: Wed Aug 22, 2012 7:10 am
Has thanked: 128 times
Been thanked: 182 times

mmmc_35 wrote: Thu Feb 08, 2024 9:25 am I really like the idea. The problem with any stat is it is like your brother in law it will lie when he has to.

Everything stat has to be taken as a gauge with your eyes evaluation.

PFF stats. People hate on PFF. I think when you see a player, you gauge his play but its subjective. PFF can bring you closer to reality. PFF is also subjective.

Let's take gervon Dexter. My eyes tell me hes a force. He progressed. I believe he played well. PFF says he played poorly. We can make excuses for the grade. I saw a young player really developing. He played very well at times. I give him a B grade. PFF grades him like a D.
last year he probably therefore a C player. With upside.

I think we all viewed the overall offense with some holes. Mooney who I like didnt play great. The center play was poor. Play calling poor. Individually I think they have some pieces. I think this stat kind of verifies that sentiment.
I think play calling is a huge part of this. All these stats begin after the play call so if that was awful then it just isn't taken into account at all.
"I wouldn't take him for a conditional 7th. His next contract will pay him more than he could possibly contribute.".

Noted Brain Genius Malk, Summer 2018.

(2020 update, wait, was I right...)
User avatar
southdakbearfan
Head Coach
Posts: 4482
Joined: Wed Jul 20, 2011 11:23 pm
Location: South Dakota
Has thanked: 623 times
Been thanked: 285 times

I have grown to despise pff, but there is a lack of statistical grading so they are kind of a default.
User avatar
mmmc_35
Hall of Famer
Posts: 6092
Joined: Sun Aug 26, 2012 12:25 am
Has thanked: 99 times
Been thanked: 83 times

malk wrote: Thu Feb 08, 2024 9:36 am
mmmc_35 wrote: Thu Feb 08, 2024 9:25 am I really like the idea. The problem with any stat is it is like your brother in law it will lie when he has to.

Everything stat has to be taken as a gauge with your eyes evaluation.

PFF stats. People hate on PFF. I think when you see a player, you gauge his play but its subjective. PFF can bring you closer to reality. PFF is also subjective.

Let's take gervon Dexter. My eyes tell me hes a force. He progressed. I believe he played well. PFF says he played poorly. We can make excuses for the grade. I saw a young player really developing. He played very well at times. I give him a B grade. PFF grades him like a D.
last year he probably therefore a C player. With upside.

I think we all viewed the overall offense with some holes. Mooney who I like didnt play great. The center play was poor. Play calling poor. Individually I think they have some pieces. I think this stat kind of verifies that sentiment.
I think play calling is a huge part of this. All these stats begin after the play call so if that was awful then it just isn't taken into account at all.
Yes. Look at brock purdy here. Hes got legit weapons at all stages. And his coach is known as an offensive guru. Correctly so.

Interesting too, other then a few outliers you could say its top to bottom a fairly accurate depiction of how good the QBs are. Mahomes, jackson, Stafford, goff good. Zach Wilson bad.
User avatar
wab
Mod
Posts: 29414
Joined: Sun Nov 02, 2008 12:49 pm
Has thanked: 112 times
Been thanked: 1779 times

malk wrote: Thu Feb 08, 2024 9:36 am
mmmc_35 wrote: Thu Feb 08, 2024 9:25 am I really like the idea. The problem with any stat is it is like your brother in law it will lie when he has to.

Everything stat has to be taken as a gauge with your eyes evaluation.

PFF stats. People hate on PFF. I think when you see a player, you gauge his play but its subjective. PFF can bring you closer to reality. PFF is also subjective.

Let's take gervon Dexter. My eyes tell me hes a force. He progressed. I believe he played well. PFF says he played poorly. We can make excuses for the grade. I saw a young player really developing. He played very well at times. I give him a B grade. PFF grades him like a D.
last year he probably therefore a C player. With upside.

I think we all viewed the overall offense with some holes. Mooney who I like didnt play great. The center play was poor. Play calling poor. Individually I think they have some pieces. I think this stat kind of verifies that sentiment.
I think play calling is a huge part of this. All these stats begin after the play call so if that was awful then it just isn't taken into account at all.
I have to wonder if there is some sort of quantifiable thing they use that incorporates some aspect of the playcall. Down and distance, route type/depth, pass vs run... like if it's a 4th and 2 and the OC calls a pass play with certain protections, and the QB takes a 3 step drop but it's also a long developing play kind of thing. IDK...I doubt they are going that far into it.
Magilla_Gorilla
Journeyman
Posts: 156
Joined: Thu Apr 02, 2009 3:54 pm
Has thanked: 63 times
Been thanked: 86 times

wab wrote: Thu Feb 08, 2024 11:04 am
malk wrote: Thu Feb 08, 2024 9:36 am

I think play calling is a huge part of this. All these stats begin after the play call so if that was awful then it just isn't taken into account at all.
I have to wonder if there is some sort of quantifiable thing they use that incorporates some aspect of the playcall. Down and distance, route type/depth, pass vs run... like if it's a 4th and 2 and the OC calls a pass play with certain protections, and the QB takes a 3 step drop but it's also a long developing play kind of thing. IDK...I doubt they are going that far into it.

That's probably way too subjective even for the PFF graders. Hoge talked about these type of situations when he was doing his end of season grading for Justin - there are times where you just can't grade the play because it was the third screen in a row and the DB jumped the play and Justin had no chance of throwing a successful ball. For individual graders I think you just move on to the next play, but for a 'stat' I don't think its really possible to quantify 'stupid OC'.
User avatar
dplank
Hall of Famer
Posts: 11676
Joined: Tue Nov 29, 2016 9:19 am
Has thanked: 1122 times
Been thanked: 1948 times

The way I've seen coaches grade players was pretty simple. They watch a play (important to note that they know the play, responsibilities, timing, progressions, etc - something internet sleuths don't know), and then they give it one of 4 grades, minus, no grade, plus, double plus. They consider all of the game context when they do this, so if it's 4th and 2 and the QB throws a ball to the right player, on time, but slightly behind him which causes him to slow up a touch and ultimately allow the defender to tackle him 6 inches short of the first down, that's a minus. If there's a quick bubble screen, even if it ends up going for 80 yards, that type of play is usually a "no grade" for the QB (but the OL/WRs all have grades), unless there's something more to it (like he checked to the play or he throws a bad ball). There's hundreds of variables they have in their head that ultimately ends up into this simple score metric - its all super subjective to the coach. Each position coach does these grades, not the coordinator or the head coach - each position coach grades all of their players and those reports roll uphill.
User avatar
Arkansasbear
Head Coach
Posts: 4633
Joined: Tue Jun 16, 2020 10:41 am
Has thanked: 415 times
Been thanked: 598 times

dplank wrote: Thu Feb 08, 2024 3:10 pm The way I've seen coaches grade players was pretty simple. They watch a play (important to note that they know the play, responsibilities, timing, progressions, etc - something internet sleuths don't know), and then they give it one of 4 grades, minus, no grade, plus, double plus. They consider all of the game context when they do this, so if it's 4th and 2 and the QB throws a ball to the right player, on time, but slightly behind him which causes him to slow up a touch and ultimately allow the defender to tackle him 6 inches short of the first down, that's a minus. If there's a quick bubble screen, even if it ends up going for 80 yards, that type of play is usually a "no grade" for the QB (but the OL/WRs all have grades), unless there's something more to it (like he checked to the play or he throws a bad ball). There's hundreds of variables they have in their head that ultimately ends up into this simple score metric - its all super subjective to the coach. Each position coach does these grades, not the coordinator or the head coach - each position coach grades all of their players and those reports roll uphill.
YOur point about the coaches knowing the play/responsibilities is huge in IMO. That's why everyone else grading OL play doesn't know enough to give a true grade.
Magilla_Gorilla
Journeyman
Posts: 156
Joined: Thu Apr 02, 2009 3:54 pm
Has thanked: 63 times
Been thanked: 86 times

Arkansasbear wrote: Thu Feb 08, 2024 4:14 pm YOur point about the coaches knowing the play/responsibilities is huge in IMO. That's why everyone else grading OL play doesn't know enough to give a true grade.
In regards to Hoge, an outsider, grading the plays - he spoke about this as well. Any play where there was any uncertainty on, he skipped until he was able to talk to a player/coach about it off the record. If he couldn't get an answer off the record he just skipped it and it wouldn't go into the final grade.
Post Reply