Obsession with stats

[Hoofbite]

Agreed 100%. A good example of this is the Cowboys offensive line.

Somehow Dallas has found a way to make the offensively appear to look pretty in the stat column without actually looking competent at all on the field.

Pretty cool trick.

 

[Seven]

Stats are for the excuse mongrels. There's waaayyyy to many intangibles for them to be considered legit. JMO, of course.

The 'I must be right' crowd bows to them as gospel cause they refuse to admit what their eyes and years of experiencing football has shown them.

 

[Clove]

I personally like the eye-witness stat. I like looking at things, then going back and re-watching to make sure your eyes aren't lying. I really don't care about the technical mumbo jumbo, because I can't do anything about it. I like the challenge of not knowing all the facts, but looking at a guy, and just knowing he's not getting it done, or he's getting it done.

Example, Spencer. With my own eyes, I just don't see this guy doing much, or at least getting consistent pressure on the QB, for a 1st rounder. I have no clue about his stats, nor do I care. But the word is, he's doing pretty good.

Whatever. We'll see how long his EXCLUSIVE to stopping the run keeps him here especially for being such a high pick and all. Anyways, that's my 2 cents.

 

[shockandroll]

In a way you are agreeing with me it seems. If you account for all of the situations that arise in a game and quantify them then yeah, statistics are useful. If you wanna argue who is the better QB Romo or Eli and you throw out completion percentage and QB rating then you're missing far too much.

There is obviously a reason why coaches chart games and pay attention to tendencies. The thread seemed to be discussing fans on the board using general statistics to diffuse arguments or pump up their own agendas. As far as that goes I stand by my post. If you're talking about really diving into every number and breaking it down based on situations and accounting for as many variables as possible you can paint a more accurate picture with statistics.

I do appreciate that we can have a disagreement and it doesn't devolve into "you're a moron who doesn't know football"

Well, first; you're a moron who does not know football! But seriously, I do agree. The point of the post is for us all to realize stats are only part of the argument, not all of it.

I'll use Big Ben as an example. Throw out his first SB cause he was useless then. But since then, while his pure passer stats have not always been great, what he did to keep drives alive and do enough to win was. He's no Brady, Manning or Rogers, imo. But he keeps fighting and gets wins where others would not.

He's not always pretty, but he does enough to win. He's a d**k off the field though.

 

[shockandroll]

Stats are for the excuse mongrels. There's waaayyyy to many intangibles for them to be considered legit. JMO, of course.

The 'I must be right' crowd bows to them as gospel cause they refuse to admit what their eyes and years of experiencing football has shown them.

Yeah, you missed the point of this thread, move along.

 

[shockandroll]

I personally like the eye-witness stat. I like looking at things, then going back and re-watching to make sure your eyes aren't lying. I really don't care about the technical mumbo jumbo, because I can't do anything about it. I like the challenge of not knowing all the facts, but looking at a guy, and just knowing he's not getting it done, or he's getting it done.

Example, Spencer. With my own eyes, I just don't see this guy doing much, or at least getting consistent pressure on the QB, for a 1st rounder. I have no clue about his stats, nor do I care. But the word is, he's doing pretty good.

Whatever. We'll see how long his EXCLUSIVE to stopping the run keeps him here especially for being such a high pick and all. Anyways, that's my 2 cents.

I can agree with this.

 

[Hoofbite]

Somehow Dallas has found a way to make the offensive line appear to look pretty in the stat column without actually looking competent at all on the field.

Pretty cool trick.

Correction.

 

[bbgun]

Stats are for the excuse mongrels. There's waaayyyy to many intangibles for them to be considered legit. JMO, of course.

The 'I must be right' crowd bows to them as gospel cause they refuse to admit what their eyes and years of experiencing football has shown them.

72% of those surveyed agree with you.

 

[shockandroll]

Correction.

This speaks to the point of this thread directly! Statistically the line is sound but we all know the truth. As I said in another thread, if they have 12 great plays and one really bad one that leads to a turnover, they are statistically good, but also lost the game.

 

[shockandroll]

72% of those surveyed agree with you.

You're on a roll tonight! You see, stats show that 98% of your posts are negative and meaningless, but that does not negate that this was a great post!

 

[TheDude]

Another stat guy (economics/finance) here. Most stats used in football are ex-post rather than ex-ante. That is, stats frame a picture of the outcome, not the causation. That picture sometimes is obvious (i.e. Rodgers completing 72% of passes, tons of TDs, yards,wins, etc.). Somtimes they do not give you what you expect (i.e. Dallas pts per game in 2010 ranked 7th with Romo out much of the year vs 10-18th in 2008,9,11 with romo).

Football is much harder to find significant stats compared to baseball or even basketball, because it is situational (play calling) and it requires at least 6-10 people doing their assignments for the whole to be successful. There are less one-on-one battles as in baseball (i.e. batter hits lefty curves better than sliders, ob% etc.)

That said, they do help frame a picture and can support a thesis on its face

Where the better use of stats would come in is when coaches chart plays to pick up situational play tendencies in order to formulize a strategy, etc.

Additionally, you could run factor, step-wise or Principal component analysis (PCA) with vast amounts of data to pick up on significantly correlated varaibles in order to project outcomes within certain probabilites. For example (and I do not have the data). you could run 500 factors through a regression model to obtain the relevent measures that when combined could give you a formula to predict an outcome within a certain probabilty (i.e. 95%, etc). A model could spit out that only 4 variables are significant. the model provides you a coefficient or "Beta" for each factor allowing you to project an outcome via a quadratic equation, orthogonal series, etc.

Even if you achieved a 99% probability of victory if you had a qurterback rating differential of 25, +1 turnover in redzone, non-off TD, < 20 yds penalties, etc), the team still has to achieve that and the play must be called and executed to take advantage of situations and tendencies.

That's why it is near impossible to use stats like you can in baseball. And injuries, weather (which can be factor variable) and low probabilty of similar year-over-year team performance strengths.

 

[Idgit]

I think UFCRules had the best response about stats. He said they're like a bikini. They reveal a lot but cover up the important things -- something like that.

It was a great line.

It was a borrowed line, and an old one. And not really an accurate one.

Statistics are tools for quantifying for the purposes of describing an argument. They're only reliable to the extent they quantify the argument properly to begin with. And they're further limited by the validity of the argument they attempt to describe.

People complain about statistics only when they either can't refute the quantification, or can't figure our how how to interpret the data to support the argument they choose to believe.

Coming out against statistics in general is an instant invitation to skip a post. Why would anyone care what you choose to believe, when you don't know yourself why you choose to believe it?

 

[realtick]

And pie. But not together, that would be weird.

Although they could metaphorically stand for something else too...

Taco truck tacos....bangin'

 

[realtick]

It was a borrowed line, and an old one. And not really an accurate one.

Statistics are tools for quantifying for the purposes of describing an argument. They're only reliable to the extent they quantify the argument properly to begin with. And they're further limited by the validity of the argument they attempt to describe.

People complain about statistics only when they either can't refute the quantification, or can't figure our how how to interpret the data to support the argument they choose to believe.

Coming out against statistics in general is an instant invitation to skip a post. Why would anyone care what you choose to believe, when you don't know yourself why you choose to believe it?

Quoted for truthiness.

 

[Hostile]

Great thread.

I personally love stats. In particular I liked career stats and records. I literally visit those sites on a weekly basis just to look at where our guys stand on the all time lists. It gives me great pride to see Emmitt atop the rushing list every time I see it. It also thrills me to see Tony Dorsett still in the top 10.

I love it that people used to wonder if Tomlinson can catch Emmitt and now you have to wonder if he can even catch Barry. I doubt it.

Stats tell you stories. Some lie to you. Some tease you. But they are a story teller and they are fun to listen to. I actually like threads and discussions about stuff that go to stats even if people disagree, because you almost always learn a new way of looking at something.

I always thought I would love to be an NFL stats guy. In particular I'd like to be allowed to go back in time and game by game count sacks and tackles and make those records more meaningful. I would consider that a huge honor.

People say stats don't matter. That is a lie. If they didn't matter we wouldn't keep track of them. If they didn't matter no one would care about home run records, rushing records, and wins. Yes, wins are a stat too.

All stats matter to something or someone. Usually the only people who get upset about stats are people who don't like to be challenged.

The old saying is, "There are lies, damned lies, and statistics." To that I say thank heavens.

 

[Future]

The world would be a better place if the only stat anybody kept was wins and losses.

 

[Hostile]

The world would be a better place if the only stat anybody kept was wins and losses.

I totally disagree.

 

[TheDude]

Great thread.

I personally love stats. In particular I liked career stats and records. I literally visit those sites on a weekly basis just to look at where our guys stand on the all time lists. It gives me great pride to see Emmitt atop the rushing list every time I see it. It also thrills me to see Tony Dorsett still in the top 10.

I love it that people used to wonder if Tomlinson can catch Emmitt and now you have to wonder if he can even catch Barry. I doubt it.

Stats tell you stories. Some lie to you. Some tease you. But they are a story teller and they are fun to listen to. I actually like threads and discussions about stuff that go to stats even if people disagree, because you almost always learn a new way of looking at something.

I always thought I would love to be an NFL stats guy. In particular I'd like to be allowed to go back in time and game by game count sacks and tackles and make those records more meaningful. I would consider that a huge honor.

People say stats don't matter. That is a lie. If they didn't matter we wouldn't keep track of them. If they didn't matter no one would care about home run records, rushing records, and wins. Yes, wins are a stat too.

All stats matter to something or someone. Usually the only people who get upset about stats are people who don't like to be challenged.

The old saying is, "There are lies, damned lies, and statistics." To that I say thank heavens.

Stats are very important to frame a picture, and I think they are even more important on an indivdual level when rank ordering or contract time.

they do matter when trying to find related stats that paint the story. Most people missuse and misinterpret stats for a sole reason to boost their arguement. If a the last 5 SB winners are in the top 4 in ydg/game, that may have absolutely 0 statistical significance when introducing ptg/game, turnovers, penalties, etc, etc.

2 factor correlations are better, but you can find correlated data in unrelated things (e.g. the cowboys perform better under a democratic president than a republican). From an outside observer, I would live to know how NFL play calling and stats are used vs gut feel. There have been studies that say you should always go for it on 4th and < 2 yards. I haven;t seen many studies taking that to a multi situational outcome (Line of Scrimm, score, etc). Im sure you could do it with the correct amount of data, But the data needed is probably 100x the variables you need for baseball.

Anyway, great discussion, I am a strong believer that stats, economics and finance are so woefully undertaught and understood that most people lose 25-40% of their wealth on bad decisions. For instance, the Financial meltdown of 2007-8 was statistically a 1 in 10,000 year event. That is why the banks weren't prepared. All of the stats projecting market performance from 2002-6 were projecting continued growth. What went wrong was fundamental flaw that leverage and loss coverage ratios woefully underestimated (volume/law of large numbers, resets and burnout). However, the fundamental issue is a person making $40K has no business buying a $600K house and no bank should make those loans especially on that scale. Didn't need stats for that. Just like we didn't need stats to see Brad John sucked out load