Suprising Stats on Miami's Pass Attack

PFF are stat monkeys.

 

To me that's no a problem of too much info, but a problem of using the info or getting the wrong info. I mean if you wanted to figure out why bees are dying it probably wouldn't make a whole lot of sense test the processor speed of Macs.

For example, if you were measuring release speed, you can't really have too many data points (or info) especially in wanting to understand if the speed changes given certain circumstances.

 

I think the analytics we're talking about here are mostly trends

For example. When Baltimore has a small lead or deficit they usually come out in a 3 WR set on 3rd and 6+.

Or. In a close game on 2nd and 7+ Kevin gillbride almost always (77%) runs out of a 2 TE set

These analytics on trends help you prepare for every situation before the game starts. Players know as soon as a play ends what set they have to be in. When they have to be on the field etc.

The above were just examples

 

It's really useless information imo because its never going to make you aquire or not aquire a player..so what exactly are you using it for, here's an idea, the time one takes looking at stats, use it for watching the players reps.

 

I don't agree with that at all. I think stats can help you decide whether or not to acquire a player. You combine that with the film study to reach the best conclusion possible.

 

What???...BS...grading a player individually on each play on whether or not they won the matchup, are not stats..their findings, grades..

 

There's no combining, no legitimate pro would ever let a stat sway them either way, think about what your saying in the game of football, your saying numbers that have no context or variables are helping you make a decision where your job is on the line...you can bet your *** no GM is gonna watch less film because of stats.

Football stats are fun to look at, if you wanna prove that stats will help you make a personnel decision, then paint the picture?

 

“Statistics are like a bikini. What they reveal is suggestive, but what they conceal is vital.”

 

I think folks who are defending the stats in this topic should tell us exactly whet their referring to?

 

Not so much as the more data you collect, the more difficult it then becomes to create a applicable, workable, methodology to use the information.

For example, once metrics for weather are added, what then?

 

I agree, Pff will put the stat monks out of business pretty soon..of course there are some things that a + or - cannot interpret, but it's sure as hell better than anything we got going.

Not knowing their assignments is a legit point, however, watching the player win or lose the battle is not rocket science, the actual grading of every play is all I can ask for, and to me Pff is used to confirm suspicions or hunches, so much better than per game or yearly averages,

 

You can combine them. NuGap already said that the 49ers and Ravens are two teams who are heavy on statistics. I imagine they combine the statistics as well as the opinions of their scouts to come to the best conclusion possible whether it is about a player or the gameplan against the opponent.

 

I'm sure organizations within the NFL have no problem working out statistics. It's not that complicated.

 

"Hello I'm Aldon Smith. Perhaps you should get to know me."
[video=youtube;M2bTJ8pY8QM]http://www.youtube.com/watch?v=M2bTJ8pY8QM[/video]

Look at the way he abuses tackles. His bull rush is ridiculous. That one against Chicago he literally pushes the tackle past the QB in front of him and just has to turn for the sack.

 

PFF doesn't just grade players though, they put together an overwhelming amount of statistics from the games they watch. Their main function is +/- that they show to rank players, but anyone with their advanced data packages (not me), can tell you that they have a lot of data on players too. Here for instance is their "elusive ranking", which is pure statistics: https://www.profootballfocus.com/blog/2012/07/01/three-years-of-the-elusive-rating/

A few weeks ago I sat across the desk from a GM who was considering establishing a statistical department. He was very interested in some of the work recently on QBs, throws into coverage, depth of routes, accuracy on the run. As an example though, he said he valued RBs that broke tackles, gained yards after contact, etc. Thus, if he were utilizing analytics the team would create a formula (maybe similar to PFF's elusive rating, but more tailored for the team) that would rank the RBs by this metric. It then allows him to more closely watch a RB to see whether those statistics show something that he missed the first time or unearth a new prospect.

When humans observe things, we have natural biases. Maybe we come in with a preordained opinion and we see something that agrees with our opinion and we fall victim to confirmation bias. Maybe we watch a player and only remember the good or bad things (I certainly have) and don't form a rounded opinion. Statistics see every play and count the middle plays, not just the highs and the lows. They're not a replacement to scouting, but a complement. There are of course many different types of stats than the one I mentioned, but those are just the type that I'll elaborate on for now.

EDIT: Here's a quote from an article about the Jags' draft. In the Draft Forum I wrote about some underrated players right before the draft, one of them being Josh Evans from Florida due to a bunch of key statistics. A day after the draft, this article came out on NFL.com. http://www.nfl.com/news/story/0ap10...s-start-anew-with-david-caldwells-debut-draft

A key quote from it:

Teams are using statistics in player evals as well as other facets of the game. It's just how the game is going right now.

 

Problem is, coaches change over time, so the current coach whom that big investment makes sense, would not make sense for the next regime who comes in.

This is where complexity and vision comes into play.

This is also why I'm pleased with the new approach as ceterus parabus, Belicheck is likely to be in NE for at least 5 more years, ie, the expected service prior to FA that any draftee will have today.

Since roads to the playoffs run through NE, it then makes sense to not try to out Belicheck Belicheck, it makes sense to take a different approach that devalues scheme in favor of impact physical talent.

Can't coach speed, speed is also difficult to coach against as things happen on the football field

 

Sorry, those are stats Deej.

 

Personally, I don't consider those numbers stats Fin, they are conclusions based on human beings making assessments, stats are not, their just accumulative numbers regardless of circumstances and variables..

Maybe those who are saying that scouts or Gms will use stats to help them make personnel decisions can define what context their using the word?

I'm not saying it isn't beneficial to look at a qbs numbers on first down to gameplan against a defense, I'm just saying that by no means does it come into play when making a personnel decision..

 

Maybe were disagreeing with the definition of the word, the numbers PFF comes up with are based on the eyes..before PFF existed it was just cumulative numbers, I.E. "stats", maybe PFF needs to separate themselves from the word, cause the word "stat", to me is empty..A number that caculated all by itself...you know what I'm trying to say?

 

I mean on the article I linked (https://www.profootballfocus.com/blog/2012/07/01/three-years-of-the-elusive-rating/) they had a formula: [(Missed Tackles Rush + Missed Tackles Rec) / (Rushes + Receptions)] * (Yards per Carry after Contact / Att. * 100)

Functionally that's not an observation right? That's a derived statistic. Every statistic is based on some sort of observation, once you quantify those observations into an average, median, etc they then begin to count as a statistic yes? I'm not sure what else would be considered a nominal statistic if that isn't.

 

Problem with pff, one of them, is it would take a NFL vet to watch the film then decipher what happened, it's all fine that "oh the blitzer came free" only a pro could say "they came free b/c the blocking scheme required the G to pull out and block them, but the DT managed to penetrate stopping the G from picking up the blitzer"

They see things through a straw, not with knowledge of what a scheme is trying to accomplish either on O or D or ST.

 

Absolutely GM, there is room for inaccuracies, but if someone could enlighten me about something more accurate I'm all ears, my whole argument is saying the numbers they provide are much more accurate then the numbers that have been available for decades, so if some personnel people are choosing the pre PFF stats over PFF today, I'd love to see the logic behind that sh&$.

Also, I don't consider what PFF does synonymous with the the word stat.not at all..I know Fin doesn't agree, but I have yet to hear a logical correlation between what PFF does, and what we figuratively know as the word stat..

I don't believe for a second anyone is making personnel decisions based on the latter..

 

We'd have to know specifically what "stats" these personnel depts are actually using for this debate to go anywhere..there's two much of a difference from what I'm talking about to what I think your talking about.

 

I don't agree b/c you can have irrelevant data points or info about the release. You see it every year in the debates about the throwing motions of various QBs. IMO their mostly irrelevant. I want to know if its fast enough and if its consistent. Just about anything else is useless. They're mostly used to confirm a bias. So people who didn't like Kaepernick or Marino brought up info on the release to justify their position. Not now of course b/c those releases were fast enough and consistent and all the other info was just noise. If you paid attention to that noise you probably made the wrong call.

 

I'm not a fan of PFF. I think they're fine at counting snaps, but useless on anything subjective. I don't believe their rankings of players anymore than anybody else's.

 

The most accurate predictors are gambling lines. Those are based off of sophisticated math modeling.

Sent from my HTC One using Tapatalk 4 Beta

 

By modeling I mean creating algorithms that will predict outcomes. Quantitative analysis similar to what you see in financial markets.

Sportsbooks maximize their profit by making games coinflips. Lines moving means someone out there has a more accurate model than Vegas. The people with the kind of money to make very large bets are the ones with the most sophisticated modeling hardware and techniques. It's just a matter of market efficiency.

Sent from my HTC One using Tapatalk 4 Beta

 

This is really just the tiny tip of the iceberg. I'm of the opinion that quantitative analysis can be as useful in football as in anything else, especially baseball. There are a lot more events in football, and therefore a lot more data that could be recorded.

Its only a matter of time before they are tracking movements of the ball and the player. Once that happens it will be like opening the floodgates. Imagine being able to quantify exactly how inaccurate a pass is? How much separation a player has? How fast a passrusher's first step is? What exact angle a player took, and how close that was to the ideal angle?

Now take all that data and imagine how that could affect in-game decision making? What if an OC couldlook on his iPad and see WR X's average separation against specific CBs so far that game?

Sent from my HTC One using Tapatalk 4 Beta

 

For sure, I know Stanford put sensors/ accelerometers in their uniforms last season to measure acceleration, hit power, etc. Those would be very interesting to incorporate. STATS Inc has working SportVU for football like basketball that can track player and ball movements. Teams just are still getting past the "old school" mentality and are still working on incorporating basic analytics now. The future is certainly exciting and bright though.

 

I'm not communicating clearly.

Try this....there is no way to have too much data that actually pertains to what you're looking at. Of course data that doesn't show the things you're looking for is unnecessary but gets in the way.

 

There is SoortVU for football??? Are teams using it? Is the data available to the public?

Sent from my HTC One using Tapatalk 4 Beta

 

I can't seem to find the article on it right now, but I know it's still in the beginning stages. I think a few teams are testing having it installed on their fields, but it'll definitely be more widespread in the coming years. Here's the page that says they're developing it/ testing it:

http://www.stats.com/sportvu/development.asp

I thought I remembered reading something like 8 teams are currently trying it/ testing it, but I can't seem to find that now. Anyway, exciting stuff.

 

I do not agree that it is a useless stat when looking at corners.