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

[Jetsfan80]

28 minutes ago, kdels62 said:

For reference. 

 

In addition, hardly anyone here had Darnold with a 50+ % shot at busting.

The anti-numbers crowd would put Home Depot to shame.  It's full of tools.  

[dbatesman]

11 minutes ago, Wit said:

I too can make my mock draft and then out phony numbers next to the players to “predict” at a fifty percent rate who will bust. 

you should

[Wit]

2 minutes ago, dbatesman said:

you should

Ok. 
i think

lawerence- 8990

wilson 570

fields 320

lance 230

mac Wilson  23

trask -456

[CanadienJetsFan]

20 minutes ago, Jetsfan80 said:

Not on QB's you can't.  No one can without good data.  No GM, scout, analyst, fan, prognosticator, QB coach or former QB can get it right on QB's at a 50 % rate without good data.

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

[RedBeardedSavage]

2 hours ago, Irish Jet said:

 

Interesting, I'll have to take a deep-dive on this later today. Thanks for posting. 

[dbatesman]

6 minutes ago, Wit said:

Ok. 
i think

lawerence- 8990

wilson 570

fields 320

lance 230

mac Wilson  23

trask -456

how do these numbers break down

[RutgersJetFan]

55 minutes ago, kdels62 said:

For reference. 

It's good bringing this into the mix because it illustrates how trading down from 2 may be a colossally dumb move this year. Wilson's season checks out. His numbers check out. His pro day checks out. In a non-Lawrence year he's the #1 pick. If the medical checks out the evidence seems to be pretty clear that you absolutely roll the dice on that guy.

[kdels62]

7 minutes ago, RutgersJetFan said:

It's good bringing this into the mix because it illustrates how trading down from 2 may be a colossally dumb move this year. Wilson's season checks out. His numbers check out. His pro day checks out. In a non-Lawrence year he's the #1 pick. If the medical checks out the evidence seems to be pretty clear that you absolutely roll the dice on that guy.

Exactly. Consensus says he’s a top 5 pick. Counting stats have him as a top 5 pick. According to PFF he put up an all-time efficient season. The game tape has him as a top 5 guy. The analytics put him firmly as a top QB prospect. He doesn’t have any questions about his physical ability.There shouldn’t be a big question here. The Jets are in position to take a player they’d give up 3 first rounders and 3 second rounders for, just do the obvious thing. 

[Jetsfan80]

26 minutes 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

And since no one can predict that, why not listen to the data when its right more often than anyone else is?

If the Jets had listened to the simple data provided in this thread, they never would have drafted Mark Sanchez, Christian Hackenberg or perhaps even Sam Darnold.  And perhaps they would have taken Pat Mahomes over Jamal Adams.  Or, prior to that, Russell Wilson in the 2nd round in 2012 when Terry Bradway was supposedly "pounding the table" for him.

[Jetsfan80]

46 minutes ago, Wit said:

I too can make my mock draft and then out phony numbers next to the players to “predict” at a fifty percent rate who will bust. 

I'm 100 % positive you cannot.  

[slimjasi]

6 minutes ago, Jetsfan80 said:

And since no one can predict that, why not listen to the data when its right more often than anyone else is?

If the Jets had listened to the simple data provided in this thread, they never would have drafted Mark Sanchez, Christian Hackenberg or perhaps even Sam Darnold.  And perhaps they would have taken Pat Mahomes over Jamal Adams.  Or, prior to that, Russell Wilson in the 2nd round in 2012 when Terry Bradway was supposedly "pounding the table" for him.

Question - Does this change your outlook at all with respect to preferring Fields over Wilson? Wilson and Lawrence seem to be in a QBASE tier by themselves. 

[Jetsfan80]

16 minutes ago, slimjasi said:

Question - Does this change your outlook at all with respect to preferring Fields over Wilson? Wilson and Lawrence seem to be in a QBASE tier by themselves. 

It certainly doesn't hurt in making me feel better about Wilson.  They both have fairly high bust chances, however.  

[JetFreak89]

13 minutes ago, slimjasi said:

Question - Does this change your outlook at all with respect to preferring Fields over Wilson? Wilson and Lawrence seem to be in a QBASE tier by themselves. 

I was a pretty staunch Fields guy a few weeks ago. I hate 1 year wonders and feel like people who dismissed the popular can't miss prospect in the past have led to situations like Deshaun Watson falling down the draft boards. But I've turned my thinking around after digging deeper into Wilson and fully expect/hope that Wilson will be the pick. I'm a fan of QBASE and this definitely solidifies my new thinking. 

[jetstream23]

1 hour ago, Jetsfan80 said:

 

Almost no one here thought Rosen would bust.  Many here thought Mahomes would bust.  

QBASE having Mahomes with a higher score than Rosen is the applicable point to be made here.  

Rosen seemed to have the highest floor in that class.  Nowhere near the best ceiling (probably about 5th or even 6th in that department) but he seemed like no worse than a very high-end backup in the NFL.  It kinda feels like Mac Jones is the Josh Rosen of this class to me.

[BroadwayJoe12]

43 minutes ago, RutgersJetFan said:

It's good bringing this into the mix because it illustrates how trading down from 2 may be a colossally dumb move this year. Wilson's season checks out. His numbers check out. His pro day checks out. In a non-Lawrence year he's the #1 pick. If the medical checks out the evidence seems to be pretty clear that you absolutely roll the dice on that guy.

But why male models?

[RutgersJetFan]

3 minutes ago, Jetsfan80 said:

It certainly doesn't hurt in making me feel better about Wilson.  They both have fairly high bust chances, however.  

Fields' numbers are actually not that bad. Certainly worthy of rolling a top ten pick on. However if you're going straight off all the boxes that have to be checked, Wilson fits. If it wasn't Lawrence's year Wilson is absolutely worthy of a #1 overall selection, and probably would be pending the medicals. Fields not so much. 

[RutgersJetFan]

1 minute ago, BroadwayJoe12 said:

But why male models?

Pretty much. Subtract one's self from the analysis and go only by the boxes that have to be checked. He literally hits every note. SOS's relationship with success is anecdotal at best and if having a stupid chubby face meant anything, John Elway would be working for Amazon right now.

[dbatesman]

30 minutes ago, Jetsfan80 said:

I'm 100 % positive you cannot.  

Joke's on you nerdlinger: he already did

[Shockwave]

14 minutes ago, slimjasi said:

Question - Does this change your outlook at all with respect to preferring Fields over Wilson? Wilson and Lawrence seem to be in a QBASE tier by themselves. 

I like Fields and yes this does carry some weight. 

Zach Wilson has a 44% chance to become Upper Tier or Elite QB. His chances of busting are only 29%. Right there how does one pass on that?

What I don't like is a players draft pick has some projection about what his stats may be. For example Justin Fields has a .26 TDYAR which is really bad. Like Tim Tebow, Colt Mccoy and Kellen Clemens bad. But if hes drafted by SF he then has a .53 which is much better (Colin Kapernick/Vince Young). 

What I will say is It is certainly nice to see PFF, Football outsiders as well as NFL FO gushing over Wilson in which Statistics and their eyes match up. Certainly is making me feel better about him being the eventual pick as per leaks. 

 

[jvill 51]

Saw this earlier today but didn’t have time to make a thread. One thing I’d like to see is how the numbers compare if you take out the expected draft position of the formula. For example if Wilson goes at 2 his numbers jump To comparable at Lawrence going 1st overall, and if Fields goes 3rd his numbers jump as well. Seeing the numbers independent of the Scouts Inc rankings (which are a placeholder until the actual draft position is known) would be helpful. I also wonder how heavily mobility factors into the rankings. Even without that though, reassuring that Wilson and Fields both fared well.

[jvill 51]

45 minutes ago, Shockwave said:

I like Fields and yes this does carry some weight. 

Zach Wilson has a 44% chance to become Upper Tier or Elite QB. His chances of busting are only 29%. Right there how does one pass on that?

What I don't like is a players draft pick has some projection about what his stats may be. For example Justin Fields has a .26 TDYAR which is really bad. Like Tim Tebow, Colt Mccoy and Kellen Clemens bad. But if hes drafted by SF he then has a .53 which is much better (Colin Kapernick/Vince Young). 

What I will say is It is certainly nice to see PFF, Football outsiders as well as NFL FO gushing over Wilson in which Statistics and their eyes match up. Certainly is making me feel better about him being the eventual pick as per leaks. 

 

You beat me to it. I get why they factor in draft position, as it’s supposed to represent the view of NFL scouts on the player which isn’t easily quantified and frankly not well known at this stage. But the numbers now, before the draft, are based on the overall rankings of non-NFL scouts. It’s clear Fields would still be below Wilson (Fields’ number if he was picked 3rd is lower than Wilson’s as the 5th pick), but the gap would certainly be smaller if they get picked one after the other.

[Embrace the Suck]

The most effective predictive tool for QB's?

That would be Chris Simms.

 

[BroadwayRay]

What are the factors that result in the drop off from Lawrence/Wilson to Fields? It doesn’t compute. 

[Jetsfan80]

4 minutes ago, BroadwayRay said:

What are the factors that result in the drop off from Lawrence/Wilson to Fields? It doesn’t compute. 

There are numerous factors, but one of the top factors for QBASE, if not THE top factor, consistently has been # of games started.  Lawrence and Wilson having 3 seasons under their belts, vs 2 full seasons for Fields, undoubtedly gives them a leg up over Fields in their formula.

[JohnnyLV]

Not a gigantic fan of these ratings but they do seem to be decently strong at predicting flops. With Wlson at only a 30% flop rate that is a very good statistical checkpoint for taking Wilson 2nd. I wonder if any NFL teams take note of this. Or if they look at what Simms says because he has been so accurate.

[Irish Jet]

3 minutes ago, Jetsfan80 said:

There are numerous factors, but one of the top factors for QBASE, if not THE top factor, consistently has been # of games started.  Lawrence and Wilson having 3 seasons under their belts, vs 2 full seasons for Fields, undoubtedly gives them a leg up over Fields in their formula.

What really stands out to me from this is Trey Lance would have been off the f*cking charts if he’d had two more years even close to his 2019. I’m genuinely shocked he graded so well.

And I think they adjusted this year to look more favourably on mobile QB’s. I think that’s helped both Lance and Fields.

[Jetsfan80]

1 minute ago, johnnysd said:

Not a gigantic fan of these ratings but they do seem to be decently strong at predicting flops. With Wlson at only a 30% flop rate that is a very good statistical checkpoint for taking Wilson 2nd. I wonder if any NFL teams take note of this. Or if they look at what Simms says because he has been so accurate.

Exactly.  QBASE does a great job at weeding out the guys who you should probably NOT draft.  That is highly valuable considering how often teams fall in love with guys with tons of red flags.

[Jetsfan80]

Just now, Irish Jet said:

What really stands out to me from this is Trey Lance would have been off the f*cking charts if he’d had two more years even close to his 2019. I’m genuinely shocked he graded so well.

And I think they adjusted this year to look more favourably on mobile QB’s. I think that’s helped both Lance and Fields.

Yep.  They're always adjusting their formula to try to paint an increasingly accurate picture of how "translatable" a college QB's skills are.  With how many athletic QB's there are across the NFL, this adjustment was crucial.

The other big factor they've been working with was how to adjust for the talent around the QB.  QBASE whiffed on their grade of Deshaun Watson because they put too much weight on all the talent those Clemson teams had.  

[RutgersJetFan]

13 minutes ago, BroadwayRay said:

What are the factors that result in the drop off from Lawrence/Wilson to Fields? It doesn’t compute. 

 

10 minutes ago, Jetsfan80 said:

There are numerous factors, but one of the top factors for QBASE, if not THE top factor, consistently has been # of games started.  Lawrence and Wilson having 3 seasons under their belts, vs 2 full seasons for Fields, undoubtedly gives them a leg up over Fields in their formula.

Projected draft position, completion percentage and projected draft value of their teammates separate Wilson and Fields in their model. Wilson's experience actually hurts him as they started adjusting total games started based on new criteria. Without that latter adjustment it's feasible that Wilson may have scored higher than even Lawrence. 

[RutgersJetFan]

 

1 hour ago, jvill 51 said:

You beat me to it. I get why they factor in draft position, as it’s supposed to represent the view of NFL scouts on the player which isn’t easily quantified and frankly not well known at this stage. But the numbers now, before the draft, are based on the overall rankings of non-NFL scouts. It’s clear Fields would still be below Wilson (Fields’ number if he was picked 3rd is lower than Wilson’s as the 5th pick), but the gap would certainly be smaller if they get picked one after the other.

They're just aggregating raw predictors of success. Projected draft position is a high predictor of success i.e. a guy that scouts give a first round grade will probabilistically be a starter over a guy that gets a fourth round grade. It's an effort to quantify in the qualitative factors, or the eye test as some label it.

[jvill 51]

12 minutes ago, Jetsfan80 said:

Yep.  They're always adjusting their formula to try to paint an increasingly accurate picture of how "translatable" a college QB's skills are.  With how many athletic QB's there are across the NFL, this adjustment was crucial.

The other big factor they've been working with was how to adjust for the talent around the QB.  QBASE whiffed on their grade of Deshaun Watson because they put too much weight on all the talent those Clemson teams had.  

I get including it, but I wonder how / how much it's weighted. For instance, Wilson probably would have scored high regardless, but he had 10 rushing TDs this year. I know they adjust for opponents, but I can't see him approaching that number in the pros (and he didn't in either of his other years on the same number of attempts). I hope that didn't influence the numbers too much because I just can't see that part of his game translating to the NFL.

[jvill 51]

3 minutes ago, RutgersJetFan said:

 

They're just aggregating raw predictors of success. Projected draft position is a high predictor of success i.e. a guy that scouts give a first round grade will probabilistically be a starter over a guy that gets a fourth round grade. It's an effort to quantify in the qualitative factors, or the eye test as some label it.

Yeah I understand the purpose of it and have no problem with it going in, but might be useful to also see what the numbers would be irrespective of projected draft position, before the draft actually happens.

[RutgersJetFan]

6 minutes ago, jvill 51 said:

I get including it, but I wonder how / how much it's weighted. For instance, Wilson probably would have scored high regardless, but he had 10 rushing TDs this year. I know they adjust for opponents, but I can't see him approaching that number in the pros (and he didn't in either of his other years on the same number of attempts). I hope that didn't influence the numbers too much because I just can't see that part of his game translating to the NFL.

All TD's are weighted. Rushing DYAR and total rushing attempts are part of the formula. Raw rushing TD totals no.

[RutgersJetFan]

4 minutes ago, jvill 51 said:

Yeah I understand the purpose of it and have no problem with it going in, but might be useful to also see what the numbers would be irrespective of projected draft position, before the draft actually happens.

Easy peazy. Guys like Case Keenum and Colt Brennan would rank towards the top. Trask likely top 2-3 in 2021. 

[kdels62]

Matt Miller is a hack. And the reason why these analytics mean something is because “the biggest boom or bust prospect” cliche doesn’t actually mean anything.