Colts Steelers MNF Thread 11/28/2022

[Morkle]

Do you guys think Najee is learning to run into open space after watching every RB but him play for the Steelers and produce pretty well?

Or is he just going to continue to dance into the back of the OL?

[willeyeam]

[LITT]

joe thomas has seen some high calibre quarterbacks

[skullman80]

Suddenly yinzer nation loves a former browns player.

[Troy Loney]

He knows what bad qbs look like

[MrKennethTKangaroo]

He played 10000 consecutive snaps for the Browns. Why is he even still remotely associated with the game of football after that kind of misery

[mikey]

Yeah you don't need to know about the process if the output doesn't pass the eye test.

Well, not quite, right? When developing a metric, you walk a fine line. If you have too many things that are way out in left field, yeah, you probably have a bad metric. But if you have a metric that absolutely conforms to everything you previously believed, then you also probably have a bad metric.

So far, no one has posed anything about the eye test with regard to the field. Mike White's numbers were put out there. But does that really consider everything that PFF considers...? Doesn't seem like it.

Mike White did have three TDs...the first one was to a guy who was uncovered (DB fell down). The second one, was thrown into bracketed coverage, it went through the hands of the DB underneath, the DB over the top had fallen down and was injured, and Garrett Wilson added 37 YAC to score by running by the injured players. The third one was more "legitimate".

But again, that's just very superficial talking about just TDs, which are even rarer data points...

I'm not saying PFF is gospel and all that...but dismissing it all together, seemingly blindly (or at least, 1 out of 32-sidedly), seems tough...

[robbiestoupe]

Yeah you don't need to know about the process if the output doesn't pass the eye test.

Well, not quite, right? When developing a metric, you walk a fine line. If you have too many things that are way out in left field, yeah, you probably have a bad metric. But if you have a metric that absolutely conforms to everything you previously believed, then you also probably have a bad metric.

So far, no one has posed anything about the eye test with regard to the field. Mike White's numbers were put out there. But does that really consider everything that PFF considers...? Doesn't seem like it.

Mike White did have three TDs...the first one was to a guy who was uncovered (DB fell down). The second one, was thrown into bracketed coverage, it went through the hands of the DB underneath, the DB over the top had fallen down and was injured, and Garrett Wilson added 37 YAC to score by running by the injured players. The third one was more "legitimate".

But again, that's just very superficial talking about just TDs, which are even rarer data points...

I'm not saying PFF is gospel and all that...but dismissing it all together, seemingly blindly (or at least, 1 out of 32-sidedly), seems tough...

Hey, 'm not digging into the formula that makes PFF what it is. And yes, I'm usually just watching Steeler games and taking note of how players on both teams play. You can usually tell when a guy is having a good game, and when a guy like Mike White is lucking himself into a good game stats wise.

The previous two seasons, you would see articles like "Ben Roethlisberger is actually better than most QBs" and "TJ Watt is OK" or "Devin Bush had a great game" according to PFF. Even players on the opposing team had rankings that were just...out there. Maybe they've tweaked the formula, but my gut reaction is to dismiss it altogether.

The problem is the data is used by the media ad nauseum and at best it is an untested, unrefined system.

[mikey]

My recollection from when I was interested in PFF, I seem to recall that three different people with some positional knowledge, watch each play independently, and score it based on what the player should have done, the player's technique, and the result as affected by the player (or whatever) and grade it/average it and then they compile that information into a metric...

I'm at least fairly sure it's the best publicly available data by several magnitudes...but like I said, it's been a while since I really looked into it...and it's possible that they've fallen off or tweaked the wrong way or whatever...

[dodint]

So it's entirely subjective, then? Cool.

[mikey]

Whether it's entirely or mostly, that's exactly what you want, right? Like, any old doofus can go "this QB won the most, so therefore he's the best" or "he has the highest QB rating, so he's the best" - that's a subjective interpretation of the value of a particular stat. It's all result-based, and not process-based...so I'd argue that those are of lower value considering all of the variables that go into those: wins are composed of less than half offense (much less pure QB play), ya figure. QB rating is subject to drops, receivers running wrong routes, etc.

So, the ultimate - and meow and I did this here a handful of years ago on just a surface level with the goalies and made an "adjusted GAA" based on goal timing and quality - is someone with real knowledge, applying that knowledge, grading that individual player's every move, and producing a metric from that. I mean, that's what real analytics are. It's not, like, corsi, and other fool's gold like that...this is exactly what you want if you really want to evaluate a player with the objectivity of numbers, but the enhanced knowledge that an expert's eye provides.

Again, I'm not saying this is necessarily what PFF does exactly and that all of their people are the same level of expertise...but subjective isn't a dirty word, in the right hands (eyes?) subjective is 100% the top of the line of what you want and what best describes the situation.

[meow]

What I saw from KSP was **** perfection

[faftorial]

I just emailed Canton and told them to reserve a spot for him.

[tifosi77]

My recollection from when I was interested in PFF, I seem to recall that three different people with some positional knowledge, watch each play independently, and score it based on what the player should have done, the player's technique, and the result as affected by the player (or whatever) and grade it/average it and then they compile that information into a metric...

I'm at least fairly sure it's the best publicly available data by several magnitudes...but like I said, it's been a while since I really looked into it...and it's possible that they've fallen off or tweaked the wrong way or whatever...

Unless they've changes their grading, they didn't used to account for whether or not a player did what they were supposed to do on the rep, or if the L the graded player was given was actually because another player somewhere else on the field screwed up. This is a loopy explanation and may not be wholly accurate, but it captures the spirit of the thing: An edge rusher comes in and gets a touch on the QB, but the play was a run away from that rush, so the QB didn't have the ball and the O-tackle on that edge rusher's side of the line had pulled and was 5 yards away lead blocking when the contact happened. PFF would give that O-tackle a negative grade on that rep. Goofy stuff like that.

It's posited as this be-all-end-all thing, and is given added weight because of Colinsworth's ownership and his status as a broadcaster.

[mac5155]

He's smiling because Jeff Saturday got his first loss.

[tifosi77]

His first loss...... since the one last week.

[willeyeam]

Pickens needs to catch that.

Take the Pitt glasses off. JFC

[Dickie Dunn]

Ben wearing them Pitt glasses

[nocera]

Every time I see that dude I like him less.

[LITT]

especially with that 'hair cut'

[Morkle]

Weird, I actually find myself liking Ben more when I see stuff like that.

[meow]

Yeah no