2021 NFL Season Thread Football • Page 1357

Info: Does my team have Henry where I can live on play action?

 

Is it more valuable to have Henry or Cook, Jefferson/Diggs, Thielen, Irv, and Mattison?

 

Still the latter but this is disrespectful to AJ Brown

And Julio for that matter I know he's not who he was but how does he not get listed but Mattison does

 

I highly doubt Josh Allen would have flourished in NYG though

 

Probably not but I get to enjoy it nonetheless haha

 

Pretty sure after Jimmy calls him a legend he says “Fuck the Packers” lmao

 

LOL

 

I don’t ever see anyone arguing that

that’s definitely not my argument

 

I see it all the time in football discourse everywhere

 

Idk how much say Lou has in drafting but I absolutely love the Bengals defensive roster.

 

Hasn't a big part of the Bengals defensive turnaround been due to the free agents they brought in last offseason? Anarumo's defenses were not worth mentioning in 2019 and 2020 before they got Hendrickson, Ogunjobi, Awuzie, Hilton, and a healthy Reader (I know they signed him in 2020, but he missed most/all of that season).

 

i just feel like the testing times are already “relative” enough; you can easily see every draft cycle that average speed for a WR is the 4.4 range and anyone who runs lower than that is a burner, while anyone slower is probably going to need additional catch point skills to stand out. You can easily see every draft cycle that most players jump around 29-34” and that anyone who jumps higher, or especially if an O-lineman jumps in that range, has some serious explosiveness. Etc

what’s the point of an RAS score that’s going to mislead people to think someone has once in a lifetime athletic traits, when you can more accurately be like “okay this 4.33 maybe isnt Deion Sanders / Randy Moss / Chris Johnson level low-4.2 speed but the guy is at least Brandin Cooks / TY Hilton fast”

things like QB rating are great because it truly does have an effective weighting metric for scores vs turnovers and for comp percentage vs yards per attempt. Calculating the adjusted completion percentage for drops, factoring in the number of sacks or picks that were or weren’t the QBs fault, that kinda stuff is valuable

what I don’t like is when the formula is this really complex and nebulous regression model where you can’t even see what’s being weighted or how much. Like CPOE and EPA. That’s the shit that really bugs me because for example CPOE relies on a calculated probability for the completion rate of each kind of pass - but aggregate probability randomized by a computer is NOT mathematically congruent to the actual completion percentage. Much like a move in Pokémon with 85% accuracy actually misses a lot more than 15% of the time, you can’t just generate a data set for that sort of thing and expect it to be 100% accurate. I always find that complex stats like that have weird outliers that makes me question their legitimacy.

Like, is a 70% chance of completing a post route against cover 2 still a 70% chance if you have to do it against Harrison Smith? And what about if he’s in the same exact position on the field, but his hips are flipped the wrong way / his eyes are on another receiver, does it factor that in as well? And even then you’re not considering variance - Harrison smith might still have a 40% chance of giving up that route, but he’s gonna still make the tackle 98% of the time. Trevon Diggs? That’s like a 15% chance of an INT but a 17% chance of giving up a touchdown cause he bit too hard on the pass. I feel like you’d have to score that differently too based on how much of a liability it could be if it goes wrong.

This is fair but it makes me wish they had better quality control, and judged the quality of their analysts by more than just their tenure and commitment to the company.

 

I think regression models in data science are great for operationalizing anything that actually can be fully explained / complete conclusions drawn from the numbers. It’s great for measuring large scale trends and patterns. However, it definitely gets tricky when you try to quantify and measure individual people’s contribution to a large effort, with many subjectively measured / perceived variables that influence each of their success. And I would say the same thing if this was a homework assignment in my major and I was talking about business analytics, not football lol

 

You're judging AS, not RAS. It's RELATIVE. It's not judging based solely on position, it's taking the full athletic profile into account.

 

The bengals played against a lot of bad offenses this season. It’s not funny because of the actual coaching though, it’s funny because he coached one good season for the giants and I guarantee mara thinks hiring him would be “the giants way!”. Not to mention he’s an italian guy from conservative af si and people already think mara is a racist because he’s never had a poc coach. Like, he’s not going to even interview ryans or bienemy or leftwich but you’re going to interview that guy lol

 

another Dolphins castoff

its pretty hilarious that almost every one of the coaches that have interviewed throughout the league have some ties to Miami

 

I mean, good coaches always seem to get more opportunities with other teams.

 

I get that Jimmy fails the eye test mostly because he makes at least 2-3 horrendous decisions/throws per game, but at some point the results need to get some proper weight. He kills it in the advanced metrics, the 49ers win with him and lose without him, and his teammates love him. 49ers should absolutely go w Trey as soon as they feel comfortable because he can really unlock a crazy upside this offense has but you can still do a lot worse at QB than Jimmy G.

 

Yeah like Kirk Cousins

 

And he’s fucking hot

 

I really hope the Broncos, Niners or Raiders give the Packers some good picks and player for Rodgers.

 

I very much appreciate the thoughtfulness of this response

for RAS, I almost view it as a data viz tool more than a “stat”, if that makes sense, and I find it useful (given that I am just a fan trying to digest large amounts of prospect information quickly and not an NFL decision maker) if you use it that way. The top end is “the athletic profile overall compares favorably to other prospects”, the middle tier is “dig in further as to whether this is someone near average across the board or ‘strong’ in some areas and ‘weak’ in others” (ie. an undersized LB with strong traits vs a hulking LB who may be a little less quick), and the bottom tier is “the athletic profile overall compares unfavorably to other prospects.” I generally agree that it doesn’t separate the elite from the very good particularly well, and just wouldn’t use it that way and it doesn’t bother me to not use it that way.

I like everything you said about the QB stats and don’t disagree with any of those counterpoints. Potentially useful to note when those significantly differ from actuals and see if you agree with their assessment, but not necessarily “correct” either