Football Outsiders content published by ESPN
PDF VERSION NOW DISCOUNTED OVER 30%
Click here to buy PDF version.
Click here to buy PDF version
Official Account: @fboutsiders
Scott Kacsmar: @FO_ScottKacsmar
Ben Muth: @FO_WordofMuth
Aaron Schatz: @FO_ASchatz
Vince Verhei: @FO_VVerhei
-- plus --
Bill Connelly: @SBN_BillC
J.J. Cooper: @jjcoop36
Cian Fahey: @Cianaf
Brian Fremeau: @bcfremeau
Tom Gower: @ThomasGower
Andrew Healy: @AndHealy
Rivers McCown: @RiversMcCown
Chad Peltier: @CGPeltier
Matt Waldman: @MattWaldman
Rob Weintraub: @robwein
19 Jan 2009
Learn where Kurt Warner and Larry Fitzgerald fall on the list of the best playoff performances by DYAR. (Hint: Very high.)
Posted by: Bill Barnwell on 19 Jan 2009
35 comments, Last at
23 Jan 2009, 3:26am by
Limas Sweed wasn't last. What a relief.
I'm not sure if it's just the comment that's wrong or the stats themselves, but Derrick Mason did not score a TD this week.
You might get credit in their system for an end-zone PI - I remember one of the Baltimore TDs was teed up by a pass interference call, I think it was a foul committed against Mason.
I'm not sure that if the MVP voters had to select now Warner would be the choice. Fitzgerald has been better than anybody else in the playoffs so far, I think he could steal it.
Wow, this was an awful week for wide receivers besides Curtis and GOD AMONG MEN.
Yaguar, looks like we have the same take on Mister Fitzgerald. Holy shit. GAM is about the same as the abbreviated moniker Jerry Rice's GOAT was.
Then again, it brings up all sorts of confusing relationships with Adrian Peterson. I may have to unearth old Sunday school books or call a priest: "Father, who is superior, Larry Fitzgerald or Adrian Peterson?"
and the logical next question: "If they were on the same team, could they defeat any team that had Voldemort and Darth Vader on defense?" Might have to consult with my kids on that last one.
I've heard plenty of talk about re-assigning the MVP this postseason, using hindsight available only after we know how the playoffs turn out. After all, both Mannings, Turner, Ryan, Peterson, the Giants' O-line, Pennington and even Chris Johnson all lost in their first playoff games. But I'm surprised to hear such talk from someone writing for FO. I figured that if anyone would understand small sample size, it would be FO. Sure Warner looks good right now, but even with these last three wins the Cardinals' record on the year is only 12-7, still worse than the Colts' final mark of 12-5. Yes, the Colts are sitting at home while Warner's Cardinals are in the Super Bowl, but it's not like Peyton played poorly at all against the Chargers - his team just lost an awfully close game. Meanwhile, geography is the only reason the Cardinals qualified for the playoffs at all.
As Aaron was saying in the Audibles, just because the Cardinals have won their three playoff games doesn't mean that they didn't get blown out in three of their last five games and play very poorly at times in the 2008 season. Just because Warner has had three great games this postseason doesn't mean that the previous 16 games he played suddenly no longer matter.
I don't think FO is advocating Warner for MVP because of his playoff performance, as much as they are asking if the voters for MVP would vote for him now (especially knowing how the media likes to use the hindsight you refer to - think Dirk's MVP from a few years ago).
"Curtis is great after the catch and working to stretch the field horizontally, which would make him a great complement to a No. 1 guy. He's just not that No. 1 guy himself. Remember that $5 million we mentioned a second ago? I can think of a spot the Eagles might choose to upgrade at."
There's a decent chance Terrell Owens will be available this offseason. He may not be quite a superstar anymore, but he can definitely improve their receiving corps. They should really look into bringing him in, he seems like an Eagles kind of guy.
That was a weird comment, I thought. Are we writing off DeSean Jackson for some reason?
That was a weird comment, I thought. Are we writing off DeSean Jackson for some reason? He's only a rookie and seems to have significant big play potential.
I thought DS Jackson looked great yesterday. I have not seen too much of the Eagles, but my memory of him from college was pure speedster (which often entails poor routes, questionable hands, little discipline, etc).
Yesterday he showed great concentration and poise on that tipped 62 yd TD as well as some diving comebackers that are not the speedster's bread and butter. Looks like a complete young WR to me.
Please adjust your sarcasm detector.
What sarcasm? Barnwell is saying the Eagles have no #1 reciver. S.K. is asking about DeSean Jackson. Sure, there's a snarky comment about T.O. in there, but I think you need to adjust your "intelligent comment" detector.
If Terrell Owens somehow loses his vocal chords, is banned from giving written interviews, and does not have an agent when negotiating a position with the Eagles, a return to Philadelphia might be an outside possibility.
Just like to comment on the lack of belief in the FO staff's writings by the FO staff. If you believe in DVOA as a superior tool for analysis in NFL games, then you were all right on that PHI, BAL & PIT would be in the Conference Championships. You also explained quite well the appearance of AZ in the Conference Championships. You stated exactly what AZ needed to do to win against both CAR and ATL. So when AZ does execute on a game plan you recommended and CAR/ATL do not adjust to stop them, why does this surprise you? And you have been "apologizing" for the discrepancy for the Eagles DVOA and their record all season. AZ again folowed a game plan in which they would win or lose based upon how well a "very good, yet innacurate QB" would score more than 32 points.
AZ gambled three times and won. They had a plan, were not afraid of losing big and played aggressively. With the exception of the first 27 min. of the second half of the third game, they executed and their opponents were befuddled. Yes, they were lucky that they got to play at home once and that the match -ups in the first two games were favorable. Yet, how many coaching staffs play to not lose in the NFL instead of to win and wind up not exploiting such chances.
This success belongs to the AZ coaching staff and the players who had unstoppable faith in their coaches (well except for the brief Q rant). Unfortunately, DVOA does not specifically analyze a coaching staff. In this case, it may have accounted for the AZ success. Yes, the luck of geography favored AZ; but, how few teams would have beleived and exploited.
I completely agree with this - the whining and excuses are getting very tiresome. Anyone who knows anything about sample size knows that these things happen. Freaking out in this fashion makes the FO crew look like the worst kind of statisticians: the kind that think their numbers can predict the future instead of merely suggesting probabilities.
Do any of the FO staff have degrees in statistics?
1. Nothing here:
leads me to believe this to be the case.
This is concerning, not because things like sample size and statistical significance are either ignored, poorly understood or rarely acknowledged, but because the questions that are being asked by the readers of this site and by the owners of this site require more than a basic understanding of statistical principles and methodologies to answer.
Unfortunately, more disconcerting, is I wonder how strong the motives of the owners of this site are to pursue these questions. FO and PFP appear to have become a fairly successul business, and as such, the business interests must be factored into decisions regardng how and what to model. Acknowledging that the problem is sample size means acknowledging it may be impossible to model the desired phenomena. Alternatively, developing an even more more complex model risks losing readership due to the increase in complexity. If it is the former, readership is lost b/c the site becomes pointless. In either case, Aaron can't feed his family.
That being said, I am convinced that DVOA is a robust assessment of a team's performance. I base this on the following:
2008 final DVOA
The top 6 (Weighted DVOA) AFC teams made the playoffs
4 out of 6 of the top (Weighted DVOA) NFC teams made the playoffs
Three out of the four played in the Conference Championships.
The correlation of Weighted DVOA with making (or not making) the playoffs is r = 0.65 p<0.00001. The R-sqare of 0.469 means that 47% of whatever the difference between playoff and non-playoff teams is, is accounted for by DVOA.
Aaron Schatz has an economics degree from Brown. Not quite statistics, but good enough for me. In fact, one could easily make the argument that an econ degree could be favorable over a statistics degree for sports usage.
I could also contend that a "statistics degree" isn't at all necessary for this sort of work. Nearly ANY good B.S. program should prepare you to perform these sorts of analyses. Most math/science oriented Master's or Ph. D. programs should as well. An actual "statistics degree" is only necessary for a VERY small subset of statistical work that is performed. In fact, the employers of "statisticians" typically don't even require a statistics degree. Instead, they look at the mathematics and statistics coursework you completed during college. Sound statistical practices aren't exactly rocket science.
I will however agree... you spent a LOT of money whilst earning that degree Aaron... you should mention it in your bio.
Agree. The degree is not the issue (although it wouldn't hurt the credibility factor). The issue is consistent failure to engage in sound statistical practices (rocket science or not).
I'm not sure what you're driving at... maybe if you could give some concrete examples of FO's "consistent failure to engage in sound statistical practices"?
Admittedly, I don't comb over things thoroughly, but in general, everything looks fine, and they readily admit flaws where appropriate.
If you're complaining because they're not going into enough detail, or listing p values, or telling us their complete methodology, or something silly like that...
If FO did that, they'd be foolish. They'd be throwing a bone to a *very* small percentage of their non-paying readership that would give a damn, while destroying any possible leverage they might have if an NFL team (or oddsmaker, heh!) took some serious interest in their work.
For consistent failure, see perpetuation of the Curse of 370.
Regarding p values, describing methodology, etc: that is exactly what I'm complaining about. There is no way to independently verify and or refute their findings. Furhtermore, I understand completely that FO has become a full time business and the need to protect their work. We the readers must beware (wary) of this fact.
Curse of 370, line continuity, stomps vs guts, etc. It seems the longer FO is around, and the bigger they get, the less they seem to care about whether they're following correct methodology.
Ah, but the "Curse of 370" (and the other things Rich mentioned below) are merely proposed theories. They're hardly statistical models. By definition, they're unproven conjectures.
Incidentally, we don't need to "beware (wary) of this fact". We need to be wary of ALL facts. Even supposedly neutral academically based scientific studies are consistently tainted (intentionally or accidentally), often in ways unknown to all but a small handful of experts in a very narrow area of expertise. Look at the fairly recent controversy over Bisphenol-A for a prime example. Or the global climate change debate.
I'm not saying FO is infallible. But this isn't baseball. You don't have thousands of iterations of similar events to pour over in a single season. Hell, you often don't have thousands of iterations of similar events to pour over for the entire available data set (which in itself, is remarkable small, due to the lacksadaisical record keeping by the NFL prior to the internet age).
If you're hoping for much more, from ANYone else? I'm afraid you're bound to be disappointed. The time WILL come when their findings are verified or refuted. Once Aaron is on an NFL payroll, you'll just have to look at the standings! (Hmmm, maybe THAT's why they're such Pats homers! Heh.)
Clearly, your take on how FO promotes the the Curse of 370 and other "theories" differs from my take.
Yes neutral academic studies get tainted, but as you point out, there usually is a small, narrow band of well informed individuals who can and/or are monitoring this and can independently test the assertions made by others. When conducting peer reviewed academic research, data is kept for a specified period of time and if requested by another researcher, is surrendered for independent analysis. Additionally, methodology is clearly delineated so the research itself may be repeated as well.
FO has become a business, I get it, and I agree w/ Rich, they've gotten sloppier as the business has become more successful.
"We're not so sure James wasn't actually Marion Barber wearing a mask in the first half. Where was the guy who was breaking four tackles at a time and picking up first downs earlier this year?"
For that matter, where was that guy *ever* in Indianapolis. Instead, I remember his perpetual failures at the goal line and his propensity to fumble in important situations.
LF now holds the record for most receiving yards in a single post-season (with a game to go no less). I'm curious, based on FO stats, who might be considered to have the best receiving post-season ever ... uh, since 1995. For that matter, I think an interesting article would include a bunch of bests for a single post-season: passing, rushing, total offensive value, defense, special teams.
Anyone else like to see that?
Sure, that would be fun, especially since we have 2 weeks to fill.
(Formerly "The McNabb Bowl Game Anomaly")
I wonder what Ben's DYAR would have been if Santonio doesn't lose the ball going across the goal line and Limas doesn't drop the big TD at the and of the half.
Don't forget Willie Parker turning his shoulders too soon on another pass that could have gone for big yards or a touchdown.
Personally, I believe that Arizona played differently on the second half of the season because they could. They could have played harder and risked more injuries, but why? They had a playoff spot, so why not increase their chances by trying to win a Playoff rather than doing better in the second half of the playoffs?
There are two parts to the NFL. Qualify for the tournament (playoffs) and win the tournament. Meanwhile, the College Season depends more upon every game.
Actually, the college game works pretty much the same way. The difference is that in the pro game, teams qualify by winning games and/or divisions; in college, teams that have been selected in the preseason as contenders qualify by not losing games. Any NFL team can make the tournament, but only a small group of college teams can make the BCS game.
Only if by "small group" you mean "a team from a Big Six conference." You don't have to be ranked high in the preseason to make it to the BCS championship game - Texas and Alabama were right on the doorstep despite poor preseason rankings.
The Vikings need offensive line help, while the Bears, Lions, and Packers have significant defensive concerns.
See All XP | NFL XP | College XP
© Football Outsiders, Inc. // Site powered by Stein-Wein // Partner of USA TODAY Sports Digital Properties