Mike McCarthy's Analytics Fraud

Tommy

Well-Known Member
Messages
2,211
Reaction score
2,970
Look. I'm an engineer by training and now work at a big bank in investments. So yeah I like math and "analytics." I think they are instructive especially when careful study shows where what you believe to be true is in fact false. People who understand these counter-intuitive truisms claim a consistent edge.

There was much talk about how much Mike has learned in his year hiatus and how analytics was a big part of the thing he leaned on and has incorporated. And good lord he is not off to a great start. In fact it seems he is a fraud.

See below:



This is an awful quote. What he says here is the exact opposite finding from the analytics community.

In fact this is something I looked into when studying Zeke's pedestrian efficiency numbers last year (As a reminder we greatly overpaid Zeke after greatly over drafting him). But I digress....I posted this in August of last year...



What this shows is that play action passing is the most effective passes a QB can throw. And it hardly matters if you run the ball a lot. Or if you run the ball a little bit. And it does not matter if you run the ball well. Or if you run the ball poorly.

The point? If McCarthy is getting this basic thing wrong what confidence do we have that he actually learned anything in the last year, analytics or otherwise? And just like the Garrett era we can expect to give away the small edges that smart coaches understand and benefit from.

And that's a shame.

Well, it was a really fun week at least.

How many years do you think Jerry is going to put up with this before we get a “real” coach?
 

gjkoeppen

Well-Known Member
Messages
7,703
Reaction score
3,327
What part of rushing attempts hardly matters and rushing efficiency doesn't matter to play action success don't you understand?

What don't you get that good rushing teams have the defense playing closer to the line of scrimmage making play action much more effective. Like I said find a new hobby because it's greatly obvious you are clueless on NFL football.
.
 

kumizi

Well-Known Member
Messages
4,209
Reaction score
5,610
He hasn't coached a single game with Dallas and people already don't like him. Goodness.
Wait...is our default feeling supposed to be "we love the new Cowboys head coach hire" or should we have an honest opinion about every decision they make?

giphy.gif
 

kumizi

Well-Known Member
Messages
4,209
Reaction score
5,610
What don't you get that good rushing teams have the defense playing closer to the line of scrimmage making play action much more effective. Like I said find a new hobby because it's greatly obvious you are clueless on NFL football.
.
Just because you keep saying this won't make it true.

Literally the point of this thread is that by the numbers/analytics, this isn't true.
 

Runwildboys

Confused about stuff
Messages
50,237
Reaction score
94,107
CowboysZone DIEHARD Fan
In this case punching someone is running the ball. Something McCarthy does 40% historically.

And whether he runs the ball well or frequently the play action game has proven to be successful. The exact opposite of his quote in the OP.
I knew what your analogy was referring to, and I agree with your assessment that play action can be successful, even when you only run 40%, but when you get right down to it, 40% isn't too far from being half the time. What's important is how well you sell the run on play action. When you watch Lamar Jackson, even on TV it's difficult to tell if he handed it off or not, while I've seen others who might as well be holding the ball up over their heads. Either way, the fake causes hesitation, but in Jackson's case it very often causes complete misdirection, which is part of the reason he can run it so effectively himself.
 

Toruk_Makto

Well-Known Member
Messages
14,231
Reaction score
17,331
From your article:


If you are going to do a deduction in general you need to show how your example excludes the one you're arguing against.

The actual analysis is the entire article where he points out cases like the lowest rushing teams and particular players where the outcome is different. IOW the analysis details specific cases where it is not universal.
The obvious implication from my statement is that it is weighted by frequency among the qbs. Not an average of each qbs numbers.
 

MaaS

Well-Known Member
Messages
365
Reaction score
358
McCarthy isn't an analytics guy. Not sure that many, if any, coaches are. What matters is if he listens to the insights from the analytics folks he hires and applies them correctly throughout the game. That is yet to be determined.

It is encouraging that he at least acknowledges it matters. A huge step up from Garrett's "just beat the man in front of you" philosophy."
 

kumizi

Well-Known Member
Messages
4,209
Reaction score
5,610
Yeah, let’s just have our head coach come out and completely give away what we’re planning to do offensively on a weekly basis. Good lord.
This is a hilarious reach.

Coach makes a vague generalization that shows he doesn't understand analytics.

Homer fans: He's doing to cause he doesn't want to give away our strategy.

LOL WUT
 

Toruk_Makto

Well-Known Member
Messages
14,231
Reaction score
17,331
So you are saying thee PFF article you cited is not trustworthy? Thanks for undermining your entire argument.

This show PA did not net a benefit for everyone ie not universal:

Barrett-17.png
Yes for 4 of 28 passers the trend did not hold.

This is what we call exceptions that prove the rule.

Thinking you're the 1 of 7 that this sample size did not hold for is the fallacy of i'm different than everyone else which leads to mistakes. As I have tirelessly argued.
 

gjkoeppen

Well-Known Member
Messages
7,703
Reaction score
3,327
Just because you keep saying this won't make it true.

Literally the point of this thread is that by the numbers/analytics, this isn't true.

Gee fuzzylumkin post #87 put some numbers that were copied from obviously a different site than yours that just blow you foolish idea all away.
.
 

kumizi

Well-Known Member
Messages
4,209
Reaction score
5,610
Gee fuzzylumkin post #87 put some numbers that were copied from obviously a different site than yours that just blow you foolish idea all away.
.
You might want to read post #87 again. It shows that ~90% of the QBs in the NFL perform better when using play action. That is something we all agree on.

It shows nothing about good rushing teams being more successful on play action.

The disagreement is "you gotta be able to run the ball well to set up play action."
 

Soth

Well-Known Member
Messages
1,077
Reaction score
922
Rushing attempts hardly matters. Rushing effectiveness doesn't matter.

Taking a small Dak sample size or a few outliers doesn't disprove that correlation. Often times it confirms it.

In any case McCarthy is saying something demonstrably untrue.

Do you have the data behind the article? This is not attempt to undermine your analysis. I really want to understand why you are so confident given the lack of information on the article. For example, the article states:

"Over the past three seasons, passer rating had a 0.14 positive correlation to play-action pass attempts in games. A quarterback’s play-action passer rating was slightly impacted by the frequency of rushing attempts within a game (a 0.14 correlation), while yards per carry had a negligible effect (-0.02)."

They argue play action is important, but it "only" has a correlation of 0.14 with passer rating. Said another way, running more play-action hardly matters? Also, they say passer rating was "slightly impacted" by rushing attempts but we know correlation numbers are not causation. I am not arguing between causation vs correlation, my point is that the article uses incorrect terms, selective metrics and inconsistent conclusions.

I mentioned Dak because the same article that claims rushing attempts/efficiency hardly matters, then claims low play-action passer rating may be due to low attempts/efficiency.

I think we would need to get the full set of data they are looking at. They only disclosed passer rating in PA vs non-PA and that's it. The data behind the the running game correlation stat was not disclosed.
 

DOUBLE WING

Well-Known Member
Messages
5,658
Reaction score
5,208
There are two key reasons this thread is stupid:

1.) McCarthy never claimed to be an analytics guru. He said he spent a day with PFF to learn more, and that he plans to incorporate an analytics department to help the organization make more informed, analytical-based decisions. We should hold him to that. What we should not hold him to, is being an expert in all-things advanced statistics on day one of his hire. His job isn't to study the in's and out's of all the deep-dive data, his job is to put people in place who do and listen to them.

2.) All this whining is over a simple quote, that if you trace it back to its origin, it is really a cat-chasing-its-tail game with the DFW media, and the sole reason this is even a debate is because the major critique of him in the past was that he abandoned the run and threw the ball way too much. The origin of McCarthy's "run to set up the pass" quote was at the introductory press conference, where someone asked something to the effect of "should fans be concerned that Dallas has the highest-paid RB in the league, and you're known for being a past-first coach?" and McCarthy gave some sort of quote about leaning on the running game, likely to assuage some of those concerns. Then, as a follow-up, Jane Slater asked him about the running game quote and how it runs counter-productive analytics and he gave the canned answer we got in the OP. There's no reason to try to read between the lines on anything here.
 

FuzzyLumpkins

The Boognish
Messages
35,684
Reaction score
27,237
The obvious implication from my statement is that it is weighted by frequency among the qbs. Not an average of each qbs numbers.

There is no obvious implication. Your OP did not differentiate between QBs. The breakdown was of various biases within the average. It was still derived from the average.

When the breakdown was of the individual QBs your conclusion broke down. Given the mechanics of how teams gameplan against offenses and how defenders react within a play it makes sense.

At the end of the day, the key is getting defenders to react to run looks instead of staying back or continuing to rush the QB.
 

Soth

Well-Known Member
Messages
1,077
Reaction score
922
Yes for 4 of 28 passers the trend did not hold.

This is what we call exceptions that prove the rule.

Thinking you're the 1 of 7 that this sample size did not hold for is the fallacy of i'm different than everyone else which leads to mistakes. As I have tirelessly argued.

"Fallacy"
"Demonstrably untrue"

As a side note, you said you are an investment banker, but you sound more like a philosopher. I have nothing against philosophers this is just an observation.
 
Top