Numb-ers - The Name Just Says it All

by Elgar on Jun 25, 2009 9:03 PM CDT Comment 9 comments

Why Favre's quarterbacking stats do not carry the day in the real world.

Statistics appear in many discussions of sports. Let's not be fooled. The next time someone unloads stats on you in a sports argument, the answer is: So what? You cannot determine from stats if Brett Favre will be the best choice at quarterback or not.

Weaknesses exist in using statistics or applying any mathematical modeling. Maybe the stuff sounds like rocket science, but it's not taking you to the moon.

If the mathematical modeling of the economy had been correct about the underpinning of those agreements called financial derivatives, the ecomony would still be running like a top, but it collapsed. Why did that happen? After all, they paid big bucks and hired the best math majors to work it all out. Oops! Real people don't buy and abandon houses like the model said. Brett Favre is not an equation. He is not a record book.

Sure, a computer weaker than your laptop can navigate to the moon, but all the hardware at NSA won't guarantee who is the best choice at quarterback.

If the richest financial corporations and math geeks cannot predict the safest bet in the marketplace, how is that guy with a beer next to you going to prove Brett Favre will be a bust in Minnesota in 2009?

He's not. Let's talk.

Let me call my first witness to tell us a root of the problem. (Calling the dead for testimony is the last refuge of a scoundrel. They can't do this kind of thing on Law and Order, but being a sports blogger is more fun than being a TV lawyer.) In the case against statistics, the prosecution calls to the stand the late Alfred Korzybski, Polish-American scientist and philosopher.

The map is not the territory. - A. Korzybski

You cannot always build a mathematical model that will guarantee you what the best decision is in certain complex situations, namely those that exceed human comprehension. No computer is going to reliably tell you that a Joe Namath is going to win the Super Bowl, because even the best Madden is not the real pro football, and Joe has more variables than you know what to do with. Many games are determined by things so unusual, you'll never know what hit you.

The next time someone hands you stats, please take it with a grain of salt. Even if stats rule the past, the past is over. What happened yesterday is not a guarantee for tomorrow.

But worse than that, the relevant stats often mentioned have nothing to do with how well Favre throws. To learn more, on which stats will count in making Favre quarterback, read on

This mathematical stuff of IBM-you-make-the-Call works for picking moves in games like Tic-Tac-Toe, but if you have two properly programmed computers duking it out at Tic-Tac-Toe, they'll play to a draw every darn time until the power goes out. What good does that do us? More importantly, real life or even NFL football is more complex than Tic-Tac-Toe. If Madden 2009 were an accurate model of pro football, we'd still notice the same team does not always win the game. The best edge you have in Madden is a tendency. But which tendencies rule?

To plagiarize a line from the Bard of Avon's Hamlet, there are more things in heaven and earth, Horatio, than are dreamt of in your philosophy. A stinking Super Madden cannot tell me if Brett Favre is the best choice for a Vikings starting quarterback or not. In 1969, stats and bookies said Joe Namath was nuttier than a fruitcake, and that was before they knew the guy would wear panty hose. Joe got the Super Bowl ring anyway.

We all know those numbers that get stuck into Madden every year are not a gift from the gods, or even based on the record books, and the record books would only tell us what did happen, not what will happen. Sports ability is not measureable by individual stats. Football is a team sport. Favre alone will not determine Favre's performance. The interactions are vast. You've got no data on Favre throwing to Percy Harvin. You would need a statisically large enough sample size of them to measure the expected result of each event with any given statistical confidence level, and like it or not, Brett Favre is dynamic, meaning he changes and he's not your Packer Backer's Brett. He's not the same Brett Favre as the 1991 rookie or the 1997 Super Bowl Champion or even the 2008 New York Jet. Your history is based on a Brett Favre who no longer exists with a team for whom he no longer plays. They only give some indication of who Brett may be in 2009, but they are not the real Brett Favre.

For my next witness, I call on a dead comedian:

I went to a fight and a hockey game broke out. - Rodney Dangerfield

I loved the 1986 movie "Back to School." Dangerfield plays a rich businessman, Thorton Melon, who goes to college. He hires Kurt Vonnegut to write his literature paper about Kurt Vonnegut. Melon gets an "F" on his essay, returned with the comment from the teacher that the author clearly knows nothing about Kurt Vonnegut.

We cannot predict the future any better than we predict the weather. Things we never considered could happen do happen, and things that happen are determined by factors to small for us to measure in advance. If the meteor streaking toward us is out of range of our detectors, we face the same bad day as the dinosaurs did 65 million years ago. It does not matter if that meteor was the ultimate cause of death or not. It was not a good day to be on planet Earth.

Ah, you may say, but what has this to do with Brett Favre? We know Brett Favre, don't we? The guys been in the NFL for eighteen seasons. Certainly we can consult some sort of actuary and determine how dead his arm should be, pull a few key stats, and there you have it, no? Um, no.

The map is not the territory, and the numbers we pull are like apples and oranges. Brett's right arm is a population of one. We might assume that Brett's arm is like every other quarterback's arm, but assumptions are your worst nightmare. We already know Brett's arm is not like every other quarterback's arm. The guy has thrown with bad mechanics is entire career. Kids may say they want to throw like Brett Favre. Good coaches with tell them, no, you don't. Bad mechanics mean he's a lousy quarterback, right? (We'll discuss that in Canton, Ohio, later.)

Bubble bees fly even though aeronautical science once said they couldn't. The bees ignorantly kept airborne. Aeronautical science was wrong. We did not know enough to model and understand how those bees fly.

Numbers are like the serving suggestion on the cereal box. They're an interesting idea, but that milk, and berries, and the fancy bowl and spoon are all sold separately.

We don't know what the 2009 Brett Favre will do. Neither does Brett. That's why the games are played. The games are too complex to predict with more than a general range. That's why we watch: to see what happens.

We can't use statistics from Super Bowls to show how Tarvaris or Sage might do. They've never been in one.

Statistics are just about useless for the purpose of determining a winner, except for the two stats which will appear on the scoreboard when the gun sounds. If you've got the key's to Doc Brown's DeLorean, get me the video of those on YouTube from Miami. Otherwise, it's all talk.

Brett knows the strategy behind the West Coast Offense. He knows which plays force the defenses to do which things and how to counter them. How well will the arm hold up? No one knows.

We don't have artillery computers to make the call, We can only rely on Kentucky windage.

Throw the numbers out the window. Winning the blog agrument will not change the results of the 2009 NFL season.

You're not talking to Brad Childress or Darrell Bevell or Rick Spielman.

Tarvaris Jackson at quarterback is not likely to fill as many seats as Brett Favre in 2009. Making money is part of winning for Zygi Wilf. Only making money will keep the Vikings in Minnesota. That's the key stat.

So, yes. It's a no brainer. Barring collision with a meteor, Favre starts in purple in 2009, no matter what other stats may say.

How many throws will a quarterback throw, before he's forever benched? The answer my friend is blowing in the wind.

Whatever that means, it sells a lot more people than do Pearon's chi square tests to determine if the frequency distribution of certain observed events in a collected sample is consistent with a particular theoretical statistical distribution, where the events being considered must be mutually exclusive and have total probability.

Remember financial analysts rule the world, right or wrong.

Computers may run the trains in D.C., but I feel another song coming on.

Your statistics give me a thrill, but your statistics won't pay my bills. Now give me money. That's what I want.

Where's a Vikings cherrleader when you need one?

Fill those seats! Fill those seats! Favre's got treats, he'll fill those seats!

The Outer Limits is over. We now return control of your monitor to you. We know you have a choice in travel, and we thank you for flying with us.

This FanPost was created by a registered user of The Daily Norseman, and does not necessarily reflect the views of the staff of the site. However, since this is a community, that view is no less important.

0 recs | Comment 9 comments

Story-email Email Printer Print

Comments

How true

Elgar, good stuff. Good reads. Did you know that Big Ben was 12th in the AFC in passer rating. 12TH!! How did that work out for them Steelers?

by c.carterhof on Jun 25, 2009 10:04 PM CDT reply actions 0 recs

I would agree

But I got lost around paragraph 30. Just kidding.

I’m not a big stats guy (as I’ve said before). They rarely tell the whole story. You can usually prove or disprove a point with the same body of stats.

In football there are so many things that matter to winning and losing and over half of them can not be statistically measured. If you want to know how players play, you have to watch them. And highlights alone can be misleading.

As much as I love pro football and fantasy football, I watch every Vikings game, every play. I will also generally watch 2 other games live. And in addition to that I watch at least 3 or 4 of those directv re-recordings that cut out everything but the plays.

I really felt when Favre was traded to the Jets, that it was going to be a one year thing and that he would end up in Minnesota. The fact that it’s coming to pass is freaking awesome to me. Because of this, I made a point to watch a lot of Jet games last year (at least 8? maybe 11?). I also just wanted him to do well in New York so I could laugh at Rogers, Thompson, McCarthy, and all those dweebs with mold on their heads.

Favre played at a high level last year until he got hurt. I thought the Jets put too much on his back at times, but they were a winning team none the less. Hopefully Favre has learned from this and learned that if he gets hurt, he needs to let backup take over. After he got hurt, he still played OK, but when he threw over 10 yards, it often showed, so in the end it was a disaster. This was too bad, because he was on pace to not only make the playoffs, but have a better season than Rogers and the Pack.

Do his stats back this up? I suppose you could present them in a manner to do that, but you could also show the opposite. Bottom line, when he was healthy, he DID play well. If the Vikes get that same level of play, they will be a more complete team.

Jackson played better than Gus last year. (Although I don’t doubt Gus had better stats) Gus was good for a game or two, but he was winning because of a lot of things. JAckson was below average at the beggining of the year. He wasn’t the reason why they lost in whole, but he didn’t do much to help them win either. At the end of the year, he played better, but really not that great. I was at the Arizona game. The defense and Peterson played their asses off and in honesty Warner choked early.

There are no stats to prove it, but with Favre last year, the Vikes would have won another 2 or 3 games. His level of play was that much better. I don’t see any reason why the same won’t hold true for this year.

BTW – Farva, if you are still reading this, good for you. Also, where is my reaction from you in regards to my imposed deadline. You know I expect one.

by newmexvike on Jun 25, 2009 10:58 PM CDT reply actions 0 recs

I confess I merely skimmed this

… but, did you just spend 5,000 words basically saying “nobody knows,” over & over?

;)

by puddnhead on Jun 28, 2009 8:34 AM CDT reply actions 0 recs

LOL!

But it was still somewhat entertaining, the season hasn’t started yet and the fever’s upon us :)

by DCPurple on Jun 29, 2009 11:41 AM CDT up reply actions 0 recs

Well...

First off, good post, nice rant. I like rants. It leaves me an opening to reply in complete rant form.

I guess I am a little confused to what exactly your main conclusion is. Is it that using stats to try to predict the future is a useless action? Or is it that Brett should be the QB because he will fill the seats with money? Because you are completely correct on one of them and completely wrong on the other, and ironically one of them is directly a result of the other.

Your reasoning is solid, but there are definite holes in your logic.

From the top – down.

“You cannot determine from stats if Brett Favre will be the best choice at quarterback or not.”

-On the contrary, I intend to explain why you have seemingly unknowingly already done this.

“If the mathematical modeling of the economy had been correct about the underpinning of those agreements called financial derivatives, the economy would still be running like a top, but it collapsed.”

- I can go along with this example, sure. First, the economy is EXTREMELY complicated. There are many, many, many, many things that have effect on the economy as a whole. Many have been set in process for decades, some we haven’t even thought of, some have bigger effects, some have smaller, some lag, some etc., etc., and these effects are all networked together in an intricate web that we are constantly experimenting with. Remember, we are an experiment. How we run our experiment, our society, our country, and our economy is based off of theory and that theory must be constantly revised in order to take into effect what we are seeing happen in reality. The best way to do this process is through forecasting and economic modeling. The best way to forecast or predict or create accurate models is by looking at data. The best way to understand this data is to use the best mathematical and statistical philosophies that we have. The reasons that things like the current collapse happen is because our models, data, and theories are not 100% complete, nor 100% accurate. For instance, Economic Models are, at their base level, based off of human behavior, and these behaviors being rational. But we all know that humans don’t always act rational, hence our models will never be perfect. But to move from this fact to the conclusion that, paraphrased, ‘the use of statistics in order to try to make the best choice about our economy is a useless pursuit’, is, well just wrong. Your making a jump.

See, I think that you chose a good example in order to visualize the fallibility of statistical forecasting, and we can relate that to football and its many complexities. But this doesn’t render stats useless, far from it.

“The next time someone hands you stats, please take it with a grain of salt. Even if stats rule the past, the past is over. What happened yesterday is not a guarantee for tomorrow.”

- Absolutely true. But, then your next sentence …

“But worse than that, the relevant stats often mentioned have nothing to do with how well Favre throws.”

- 65,127 yards. 5,720 Completions. 464 TDs. 310 ints. 469 sacks. 0 games missed because of injury. I can make a reasonable assumption based on these STATS alone that Brett Favre knows how to read a defense. He knows how to pass a football really really well. He knows where to throw that football. He has been through many sunday battles and wars. He is tested. He has made mistakes. He has been hit and keeps on going. Should I render this useless? Should I forget all of this when making a decision as to whether or not Brett Favre would be the best choice to lead this team? Should I make my decision only on things present, a 40 year old guy with a throwing arm in rehab? if we didn’t look to where ‘stats rule’, because ‘the past is over’, nobody would think twice about brett. But of course we do, the stats paint a picture, and we use the picture to try to extrapolate possibilities into the future. Of course it’s not a guarantee, but…

Ok, maybe I should stop, I could do this for a long time, but it is probably useless. I mean I feel like I am trying to explain basic philosophy of how we make decisions. So I am gonna stop now. I will just say, it is obvious that statistics are a useful tool that we use. No, not perfect. But very useful. Yes Brett Favre will fill up seats, but that is because people know of his past, of his STATS, of his performances, and will want to see if he can fulfill their expectations or predictions of how that past carries forward. Along with other reasons, yes, but if you break it down and keep breaking it down, like I was going to attempt to do, I am confident that at the heart of the assess in the seats is some kind of statistical reasoning that the fans are using. And, although, I think it would be interesting to go far down that path of thought and see where exactly it ends up, or how it gets there, I should really get back to work.

A couple more of your quotes that I must comment on, Quickly
.
“but assumptions are your worst nightmare.” Wholey wrong. Assumptions are engrained in our every thought and inner psychology. Very useful tools.

“Numbers are like the serving suggestion on the cereal box.” Nope.

“We don’t know what the 2009 Brett Favre will do. Neither does Brett. That’s why the games are played. The games are too complex to predict with more than a general range. That’s why we watch: to see what happens. " Yes, this sounds reasonable. But that general range may be quite telling. Think about betting, and lines, and stats.

“We can’t use statistics from Super Bowls to show how Tarvaris or Sage might do. They’ve never been in one. " OK. And…

“Statistics are just about useless for the purpose of determining a winner,” Wrong. Easy example, 0-16 lions at the end of last year, versus the eventual champs, 12-4 pittsburgh steelers. I give those two stats to anybody with no knowledge of the team, or even football, and I guarantee that they will put their money on the correct choice. Just by using useless for the purpose of determining a winner.

“Tarvaris Jackson at quarterback is not likely to fill as many seats as Brett Favre in 2009. Making money is part of winning for Zygi Wilf. Only making money will keep the Vikings in Minnesota. That’s the key stat. " 100% Agree. But the money comes at some level due to stats that….blah blah blah. Break it down, it will come back to fans using stats.

“Whatever that means…” Kinda what I was thinking…

by footballninja on Jun 29, 2009 11:41 AM CDT reply actions 0 recs

Wow, you had a lot of time on your hands too, heh :)

Very picky, but good points.

What comes to my mind is the “x-factor”, or perhaps, gestalt effect if you prefer. Sometimes, with the right package, the whole is significantly greater than the sum of the parts, and this might apply to both Favre and to the Vikings with Favre at the helm. It happened in 1968 with Joe Kapp, it could happen again (hopefully with a better end-result) with the Vikings. Yes, I know Joe was younger, but he didn’t have that much juice left in the tank either. 1, maybe 2 more seasons, and he was done too.

A leader is more than the ability to remember what’s going on, and to tell people what to do, or even to throw the perfect pass… there’s instinct, there’s emotional impact and response, there’s inspiration, there’s confidence both in self and the confidence that others hold you in. The ability to get a little bit more out of people than they knew they had in them. Maybe even belief…

I don’t know if Favre is going to be The Man… yeah, sure, he’s being given every opportunity to do so, and to start, but he could fall apart ability-wise, halfway through the season, or get injured, or just turn out to not be the man he once was. But beyond the numbers and the hero-worship (and anti-hero hate), there’s a leader coming to Minnesota. A true field-general, a “bloody giant”. Train-wreck or triumph, 2009 is going to be a damned hard year to take your eyes off of.

by DCPurple on Jun 29, 2009 12:07 PM CDT up reply actions 0 recs