Football Outsiders QBASE: The most effective predictive tool for QB's?

[thebuzzardman]

QBASE? More like FREEBASE.

[Bowles Movement]

Sorry   Not impressed by this at all

55 percent accuracy is barely above a coin toss

there are plenty of good Qbs that would not have been selected using this   Matt Ryan grouped with Hackenberg?

this doesnt predict that Mahomes would be a star   His point total is mid pack at best

Might as well throw darts at the names blindfolded for all the help this supplys

 

[Icer]

For such a great prospect Darnold sure had a high bust rate

[kdels62]

9 minutes ago, Bowles Movement said:

Sorry   Not impressed by this at all

55 percent accuracy is barely above a coin toss

there are plenty of good Qbs that would not have been selected using this   Matt Ryan grouped with Hackenberg?

this doesnt predict that Mahomes would be a star   His point total is mid pack at best

Might as well throw darts at the names blindfolded for all the help this supplys

 

Seeing as how we’re 0-7 (ish) since Sanchez, I’d be good with 55%.

[RedBeardedSavage]

I don't get people that are so dismissive of this type of stuff.

You can point out the outliers, but technically they've accounted for that - that's why there are %'s assigned to probabilities of outcomes for each prospect.

Regardless, this is a far more interesting read/food for thought than listening to Dan Orlovsky praise a quarterback on a two minute segment on ESPN/ESPN radio with no actual nuance or substance. 

[Jets Voice of Reason]

I agree with RJF, the strength of this in its current iteration is to use it as a screener/filter for weeding out mediocre/bad prospects who may be overvalued from a draft perspective.

There is still work to be done in terms of using this as a predictive mechanism for success. At this point, with the new formula, they have to get more data points to see how accurately it'll forecast success.

The people fixated on 50% don't understand that it's a relative comparison. It's not a "will they bust or not" as much as it is hey, 50% bust potential is higher than someone else with a 20% bust potential, indicating the risk entailed with the prospect in question.

 

[Icer]

27 minutes ago, RedBeardedSavage said:

I don't get people that are so dismissive of this type of stuff.

You can point out the outliers, but technically they've accounted for that - that's why there are %'s assigned to probabilities of outcomes for each prospect.

Regardless, this is a far more interesting read/food for thought than listening to Dan Orlovsky praise a quarterback on a two minute segment on ESPN/ESPN radio with no actual nuance or substance. 

I can't think of a QB that Orlovsky doesnt like

[RutgersJetFan]

28 minutes ago, RedBeardedSavage said:

I don't get people that are so dismissive of this type of stuff.

You can point out the outliers, but technically they've accounted for that - that's why there are %'s assigned to probabilities of outcomes for each prospect.

Regardless, this is a far more interesting read/food for thought than listening to Dan Orlovsky praise a quarterback on a two minute segment on ESPN/ESPN radio with no actual nuance or substance. 

Outliers are part of statistics and a sign of the strength of the model. Would be weird if they didn’t exist.  

[Dcat]

there are five (5)  Jets QBs in the bottom 20 of that list.  LOL  we suck

[Jets Voice of Reason]

1 minute ago, Dcat said:

there are five (5)  Jets QBs in the bottom 20 of that list.  LOL  we suck

Which is exactly why there is significant value in using a model to try to filter out high bust potential prospects, or at least value them more appropriately.

[RedBeardedSavage]

11 minutes ago, RutgersJetFan said:

Outliers are part of statistics and a sign of the strength of the model. Would be weird if they didn’t exist.  

Data Analytics whipped my a$$ in grad school. I like to study history. Simple man like stories. Simple man avoid numbers. 

The R squared of me knowing WTF I'm talking about it when it comes to this stuff is low... But I'd rather struggle through this than listen to Todd f***ing Mcshay. 

[Augustiniak]

36 minutes ago, Jets Voice of Reason said:

I agree with RJF, the strength of this in its current iteration is to use it as a screener/filter for weeding out mediocre/bad prospects who may be overvalued from a draft perspective.

There is still work to be done in terms of using this as a predictive mechanism for success. At this point, with the new formula, they have to get more data points to see how accurately it'll forecast success.

The people fixated on 50% don't understand that it's a relative comparison. It's not a "will they bust or not" as much as it is hey, 50% bust potential is higher than someone else with a 20% bust potential, indicating the risk entailed with the prospect in question.

 

and also can suggest how much more likely a qb taken in the top 5 can become very good as opposed to merely a functional starter.

[Bruce Harper]

20 hours ago, CanadienJetsFan said:

Data is very important I’ll give you that.

 But it’s what’s between the ears  that counts and you can’t measure that 

Unfortunately

This is an oversimplification because what is between the ears should show up on the field,

[PS17]

18 hours ago, Embrace the Suck said:

The most effective predictive tool for QB's?

That would be Chris Simms.

 

At 3:54: "He does have a nice set of legs on him. You know I'm a legs and ass guy, and his thighs and butt -- it's an athlete's thighs and butt." Chris Simms is a treasure. 

[RutgersJetFan]

35 minutes ago, Augustiniak said:

and also can suggest how much more likely a qb taken in the top 5 can become very good as opposed to merely a functional starter.

They actually broke this down pretty well several years ago:

https://www.footballoutsiders.com/stat-analysis/2017/nfl-draft-round-round-qb-study-1994-2016

Long story short take a QB early or take a QB late with a low risk pick. There is no such thing as mitigating risk for QB's. You either go all in or say **** it at the end if you want the best chance at succeeding. 

Worth noting this all strengthens their logic in QBASE even more with respect to including projected draft position.

[RutgersJetFan]

1 hour ago, RedBeardedSavage said:

Data Analytics whipped my a$$ in grad school. I like to study history. Simple man like stories. Simple man avoid numbers. 

The R squared of me knowing WTF I'm talking about it when it comes to this stuff is low... But I'd rather struggle through this than listen to Todd f***ing Mcshay. 

That's usually the norm I experience in my classrooms, but I think people understand stuff like this way more than they think they do. Ignore all of the explanations of variables, equations, ratios, and fractions. Anyone that can understand the basic logic of flipping a coin twice in a row and that it may come up the same twice in a row can grasp something like QBASE. Sometimes the improbable happens. 

[ZachEY]

1 hour ago, RedBeardedSavage said:

I don't get people that are so dismissive of this type of stuff.

You can point out the outliers, but technically they've accounted for that - that's why there are %'s assigned to probabilities of outcomes for each prospect.

Regardless, this is a far more interesting read/food for thought than listening to Dan Orlovsky praise a quarterback on a two minute segment on ESPN/ESPN radio with no actual nuance or substance. 

In general, people hate the notion that a formula is better than "expertise," regardless of the field.

[addage]

First, a question:  isn't it implicit in the QBASE rating that a QB who opts for the draft too soon is putting his career potential at risk?  It seems QBASE is saying that there are skills that QBs learn in the college game that are far more difficult to learn once you make the jump to the pro's.  Is that a correct interpretation?

Re: Sanchez.  When he declared for the draft, the USC coaches told him he wasn't ready.  At the time, I thought that was just sour grapes over losing a QB they thought would be their starter next season.  But, maybe they were right and Sanchez did make a bone headed move.

I also seem to remember that maybe in year 2 there were some stories about his not being very coachable,  Didn't really listen to older players and coaches.  Again, may or may not be true, but it fits into what QBASE is implying.

[RedBeardedSavage]

10 minutes ago, RutgersJetFan said:

That's usually the norm I experience in my classrooms, but I think people understand stuff like this way more than they think they do. Ignore all of the explanations of variables, equations, ratios, and fractions. Anyone that can understand the basic logic of flipping a coin twice in a row and that it may come up the same twice in a row can grasp something like QBASE. Sometimes the improbable happens. 

I look at the results of their modeling - they've got a fairly good track record. That's something a scruffy-looking-nerf-herder like me can understand.

Then I take a gander at the 1 in 5 chance their model suggests that Wilson is elite. A slightly less than 1 in 3 chance that he busts. 

Then I look at ol' Fields-y, with his 1 in 8 chance of being elite combined with his 40% chance of busting.

 I'm like "Alright, the Mormon Johnny Football gets the nod by the nerds and one of the few traditional scouts I respect (Top button Simms)."

[CanadienJetsFan]

1 hour ago, Bruce Harper said:

This is an oversimplification because what is between the ears should show up on the field,

If it were that easy then we wouldn’t be waiting for 50 years for our next FQB

[slimjasi]

14 minutes ago, addage said:

First, a question:  isn't it implicit in the QBASE rating that a QB who opts for the draft too soon is putting his career potential at risk?  It seems QBASE is saying that there are skills that QBs learn in the college game that are far more difficult to learn once you make the jump to the pro's.  Is that a correct interpretation?

Re: Sanchez.  When he declared for the draft, the USC coaches told him he wasn't ready.  At the time, I thought that was just sour grapes over losing a QB they thought would be their starter next season.  But, maybe they were right and Sanchez did make a bone headed move.

I also seem to remember that maybe in year 2 there were some stories about his not being very coachable,  Didn't really listen to older players and coaches.  Again, may or may not be true, but it fits into what QBASE is implying.

Yep, this is basically my interpretation, as well. And it jives with the conventional wisdom of the old days (not applicable anymore) that you would prefer a guy who played 4 years in college.

it turns out that experience matters - just like it does in every other profession. What a concept! 
 

 

[RutgersJetFan]

46 minutes ago, addage said:

First, a question:  isn't it implicit in the QBASE rating that a QB who opts for the draft too soon is putting his career potential at risk?  It seems QBASE is saying that there are skills that QBs learn in the college game that are far more difficult to learn once you make the jump to the pro's.  Is that a correct interpretation?

Re: Sanchez.  When he declared for the draft, the USC coaches told him he wasn't ready.  At the time, I thought that was just sour grapes over losing a QB they thought would be their starter next season.  But, maybe they were right and Sanchez did make a bone headed move.

I also seem to remember that maybe in year 2 there were some stories about his not being very coachable,  Didn't really listen to older players and coaches.  Again, may or may not be true, but it fits into what QBASE is implying.

100% absolutely. They started weighting experience to account for one-year wonders within that logic as well. There is a threshold a prospect has to meet for their year of starting experience to count to levels that impact the model. That’s why Lawrence grades out so high. Multiple years at an elite level is the #1 indicator that a guy will be good. 

[whodeawhodat]

19 hours ago, Embrace the Suck said:

The most effective predictive tool for QB's?

That would be Chris Simms.

 

tbh it moved a little listening to this.

have a lot of respect for Simms grading of QBs and his past is proven.

this is so much more than being the contrarian to TL being best in class and looking for clicks.

I was on the oline train but we must take wilson @ #2 

idc if we trade up or down with the rest of our pics, grab oline later but do not pass on this kid.

[neckdemon]

On 9/19/2020 at 1:52 PM, Jetsfan80 said:

Oh wow you win this round

wrecked it

[Wit]

23 hours ago, dbatesman said:

how do these numbers break down

Ok so it’s based on a sliding scale and the amount of carbs they consume as well as the blindness of their hair and it’s length. 

[Embrace the Suck]

3 hours ago, PS17 said:

At 3:54: "He does have a nice set of legs on him. You know I'm a legs and ass guy, and his thighs and butt -- it's an athlete's thighs and butt." Chris Simms is a treasure. 

Yeah, he's used that line a few times. When you can blend humor with useful genuine insight you're better than good at what you do. He sold me on Wilson to the point that I'm hoping the Jags don't sway in their conviction and they draft the highly overrated TL. 

[Embrace the Suck]

2 hours ago, whodeawhodat said:

tbh it moved a little listening to this.

have a lot of respect for Simms grading of QBs and his past is proven.

this is so much more than being the contrarian to TL being best in class and looking for clicks.

I was on the oline train but we must take wilson @ #2 

idc if we trade up or down with the rest of our pics, grab oline later but do not pass on this kid.

Yeah, after listening to him talk about Wilson I was sold as well. He seemed very intrigued with Wilson and dripped with enthusiasm about every part of his game. He made his arm out to be legendary, and that he actually stands above TL with a real gap there.  I wouldn't be surprised if at one point after climax his wife asks him "Who the **** is Wilson? You just screamed out "Oh, Wilson"".

[dbatesman]

2 hours ago, slimjasi said:

Yep, this is basically my interpretation, as well. And it jives with the conventional wisdom of the old days (not applicable anymore) that you would prefer a guy who played 4 years in college.

it turns out that experience matters - just like it does in every other profession. What a concept! 

This is a big part of it, but it’s also that guys with more starts generate more tape, and so their traits are easier to scout accurately