Skip to content

Wait. NFL analysts don’t factor in context?

Feb 1, 2011, 1:05 PM EDT

Divisional Playoffs - Green Bay Packers v Atlanta Falcons Getty Images

Tom Verducci — a baseball guy — has a football-related column up over at Sports Illustrated today in which he throws in some baseball spice to the NFL soup:

Baseball fans get the importance of context on performance — especially where the game is being played — going all the way back to the debate over what would have happened if Ted Williams were a Yankee taking aim at the rightfield short porch in the Bronx and Joe DiMaggio were a Red Sox swatting away at the Green Monster.

Football, however, treats teams and quarterbacks with almost no regard for context. Sure, unlike baseball, the dimensions of the playing field are uniform, but football is a very different game depending on where it is played because of the elements.

Verducci then proceeds to explain how Peyton Manning and Drew Brees are somewhat overrated due to the fact that they play in domes, and how Aaron Rodgers is underrated because he plays outside in the cold all the time. He goes on to note that Ben Roethlisberger indoor/outdoor splits aren’t as pronounced, thus Rodgers has the advantage in the Super Bowl. Which is fine as far as it goes.

But with the caveat that I read absolutely zero NFL analysis and don’t really plan to, I have to ask:  is it really the case that no one analyzes this stuff already?  I know there are football websites and blogs who have followed in the footsteps of baseball sabermetrcians. This stuff has to have been covered a decade ago, right?  It seems fairly basic even to a non-analytical, non-football guy like me.

Am I wrong about that? Is it like Verducci says and football “treats teams and quarterbacks with almost no regard for context?”  Or is that just teams and mainstream media who haven’t embraced advanced analysis like baseball teams and media have?  Or is Verducci just wrong about it all and both teams and media covering the game have factored this stuff in years ago?

The NFL is a mutli-billion business and guys are on the firing line way more on a football team then any baseball manager or coach is. I have to think, then, that folks in the NFL are well aware of this kind of thing.

  1. emac11 - Feb 1, 2011 at 1:09 PM

    I disagree, people always talk about the context in football.

  2. Lukehart80 - Feb 1, 2011 at 1:15 PM

    I feel like people talk about this kind of context when discussing what will happen in a particular game (“it’s going to be cold and windy, so both teams will have to plan on…”), but I don’t see/hear it much in mainstream/network analysis of player performance over the course of a season or a career.

  3. Mr. Jason "El Bravo" Heyward - Feb 1, 2011 at 1:21 PM

    The fact that the Pack had a undefeated home playoff record for like…ever…was a huge story. That’s what made it special when Vick beat them in Lambeau awhile back. This is just one example of why Verducci is wrong.

    • Lukehart80 - Feb 1, 2011 at 1:28 PM

      Yes, but the difference that’s being pointed out (I think) is that in baseball, smart people know not to make direct comparisons between one player’s stats and another without adjusting for park factors, league differences, etc. Is that being done in football? How do Peterson’s numbers compare with Johnson’s given the context they were accumulated in? I wouldn’t claim to know for sure, but I don’t feel like I ever see or hear that kind of stuff for the NFL.

      • nps6724 - Feb 1, 2011 at 1:37 PM

        Every player plays on the same size field every week so park factors is irrelevant. And you can’t take weather into account because A) different types of weather affect games differently (snow makes it harder to run, rain makes it harder to catch, wind makes it harder to throw, etc.), and B) you would have to adjust for way too many games due to weather conditions. If you tried to factor in all these, you’d end up with a jumbled mess of fake numbers. This works in baseball because no one park plays exactly like any other.

      • Lukehart80 - Feb 1, 2011 at 2:55 PM

        I understand that it might be a more complicated process, but that doesn’t mean it can’t be done at all. I see “total rushing yards,” “yards per game,” and “yards per carry,” but rarely (never?) anything more in-depth than that. Is that really the best that can be done?

        I haven’t looked at FootballOutsiders, maybe they have all kinds of things, I don’t know.

        It just seems defeatist and/or lazy to say nothing more COULD be done.

      • churchoftheperpetuallyoutraged - Feb 1, 2011 at 3:00 PM

        Is that being done in football? How do Peterson’s numbers compare with Johnson’s given the context they were accumulated in? I wouldn’t claim to know for sure, but I don’t feel like I ever see or hear that kind of stuff for the NFL

        Footballoutsiders.com does just this, with DVOA. The adjust for tons of variables (OL quality, defense adjusted, etc). Also, their Audibles at the Line column on Monday blows PKing’s MMQB out of the water both in information and humor/entertainment.

      • nps6724 - Feb 1, 2011 at 3:31 PM

        Because at some point you break it down so far that the result is meaningless. You end up sounding like the announcer from “Little Big League” (“He’s hitting .450 in day games in odd months on the East Coast after 8 pm”).

        Baseball is more defined (for lack of a better word) than football in this area. A batter can hit with none on, runner on 1st, runner on 2nd, runner on 3rd, bases loaded, runners on 1st and 2nd, runners on 2nd and 3rd, and runners on 1st and 3rd. That’s it. Even if you include 0 outs, 1 out, and 2 outs, you’re still looking at only 24 general situations.

        Just taking down and distance into account, there are 99 different possibilities for *each down* (from 1st-and-1 to 1st-and-99). Obviously a vast majority of those instances won’t come up, but they’re still possible. Even if you exclude everything beyond 20 yards, that’s still 80 general situations, over 3 times more than baseball.

        So how would you further break down, say, how a QB does on 3rd down? They already break it down to 3rd-and-short, 3rd-and-medium, and 3rd-and-long. Would you break it down for each situation (3rd-and-1, 3rd-and-2, 3rd-and-3, etc.)? It just gets ridiculous at that point.

        Sample size + the many different possibilities works against football for sabremetric-style analysis. For instance, Drew Brees had only 17 pass attempts on 3rd-and-short in 2010. Barely one per game. Even without breaking it down further, you can tell the sample is too small to make any claims whatsoever.

        Baseball is rigid enough for this type of analysis to take place because there’s a relatively small number of situations that can occur at any one time. That isn’t the case in football.

        Another problem is the statistics themselves. OL and most special-teamers have no stats. There’s no official stat for sacks allowed or pancakes outside of Madden. And unless a special-teamer is a returner or happens to make a tackle, they can go the full season without a single statistical mark. And a WR who is never thrown the ball would have zero stats, even if they played the whole game. And on running plays, the only offensive player who will earn any stat is the ball-carrier. So if two teams combine for 45 rushes in a game, that’s 45 times 10 offensive players will finish that play with zero stats. A LT could make a crushing block that springs a RB for a huge TD yet the RB gets the reward.

        Baseball, at its core, is a team of individuals playing a game filled with consecutive one-on-one matchups where every player involved in those matchups will be counted in some fashion. Football is the opposite.

      • spudchukar - Feb 1, 2011 at 4:20 PM

        The obvious problem is the small sample size.

      • churchoftheperpetuallyoutraged - Feb 1, 2011 at 7:55 PM

        @nps6724

        Go to the link I’ll post below, and read about DVOA (Defense-adjusted Value over Average) and DYAR (defense-adjusted Yards above Replacement). They explain how they get their metrics. They also use context of the game unlike WAR (which doesn’t use WPA in it’s calculations). So a 9 yard run on first and 10 or 2nd and 3 is far better in DVOA and DYAR than it is in 3rd and 15.

        http://www.footballoutsiders.com/info/methods

  4. nps6724 - Feb 1, 2011 at 1:23 PM

    Every football game is chock full of how particular players or teams performed in the redzone (offensively and defensively), on 3rd down, on 4th down, inside the 2-minute warning, utilizing the run, utilizing the pass, against the run, against the pass, at home, away from home, in night games, on Mondays, and so on.

    And if the guy is calling Peyton Manning and Drew Brees overrated, maybe he should stick to baseball.

  5. bpipes81 - Feb 1, 2011 at 1:24 PM

    Football isn’t subject to the same kind of statistical analysis (because it is much more of a team game than baseball is), meaning that context is taken into account more for team advantages than individual performances. Drew Brees and Peyton Manning might have certain statistical advantages of Rodgers or Roethlisberger because they play at least 1/2 their games in a dome, but NOBODY rates outdoor QBs lower just because they might throw for lower yardage or TDs.

  6. churchoftheperpetuallyoutraged - Feb 1, 2011 at 1:26 PM

    Or is that just teams and mainstream media who haven’t embraced advanced analysis like baseball teams and media have?

    This x1000. Think of all the SABR friendly websites out there for baseball: Fangraphs, BtB, THT, Tango’s site, BPro, and then all the team specific sites that are SABR friendly (RAB, Crash Burn Alley, etc). There are dozens if not hundreds out there debunking all the “commonly held beliefs” that tend to be horribly wrong.

    For football, it’s footballoutsiders.com and advancednflstats.com (haven’t checked them out, but FO is amazing). It’s a shame it’s over, but if anyone is on Twitter and is an NFL fan, you have to follow @philsimmsquotes to see how idiotic one of the supposed “best announcers in the game” is (never mind Matt Millen).

  7. Chris Fiorentino - Feb 1, 2011 at 1:28 PM

    Craig, I’ll make it very easy for you. Verducci is 100% wrong here. He should stick to baseball.

    • paperlions - Feb 1, 2011 at 1:51 PM

      +1

  8. poreef - Feb 1, 2011 at 1:51 PM

    The NFL operates as a pretty closed market in all facets of labor – from the players to the coaches and staff. It’s pretty rare for new people and ideas to creep in. But, once they do, (the Wildcat, etc) they get copy-catted to the point where no longer work. So there’s kind of a self-fulfilling conservatism that remains very strong in football.

    The football sabrmetricians have very clearly found a few things, but they’re more or less screaming into the ether. I’ll mention two: (1) blitzing a good QB is self-defeating because they don’t panic and have a hot read in mind and (2) going for it on 4th down almost always makes more sense than punting the ball away. The latter is more or less the equivalent of “don’t bunt” and the former is something like “stealing bases is overrated”.

    • frug - Feb 1, 2011 at 5:07 PM

      Actually neither of those is quite true. The usefulness of a blitz depends heavily on the down and distance (to say nothing of scheme and personnel) as well as the size of the blitz. Football Outsiders did a big analysis in one of there books a couple years that showed on average it is perfectly reasonable (and in many cases advantageous) to send a “big blitz” (7+ pass rushers) on (I believe) second and long and third and short. On the other hand mid-size blitzes (6 pass rushers) or or no blitz have there own benefits. Unlike baseball where stealing bases is overrated almost always, blitzing on average is neither good nor bad.

      Also, the never punting thing is tough to test because no coach has ever made a serious attempt at it. At this point all estimates are purely theoretical. This is to say nothing of the fact that the advantage teams get from going for it on fourth, or not, isn’t the same for every team. A team with a powerful offense and crappy defense and special teams like the Colts of the mid-decade would generally benefit from being really aggressive on fourth down, while a team like the Bears with a bad offense but excellent defense and special teams will benefit from playing the field position game.

      • poreef - Feb 1, 2011 at 5:22 PM

        The Bears couldn’t punt past 20 yards from scrimmage to save their life in the Packers game. But, leaving aside that…

        My point is two-fold and has nothing to do with the broad examples I posted to give the baseball guys on this forum a gut instinct for analogous conclusions in baseball. (1) NFL commentators (think Tom Jackson or Ditka whenever they wheel him out from his dungeon) wave around old saws about running, defense and pressure on the quarterback as though they were handed down from heaven. Modern statistical analysis suggests that there many of these conclusions are poorly based, only situationally important or even counterproductive. (2) the NFL community (coaches and ownership/organizational structure) at large is extremely risk-averse and poor at incorporating new ideas into the game.

        Many of the same things used to be said about baseball ownership. Now every team incorporates these ideas into their thinking – some better than others, of course.

        Until the NFL organizational structure opens itself up to new ideas, these analyses are outside the mainstream of the game itself and fan base generally. It’s not called Football Insiders.

      • nps6724 - Feb 1, 2011 at 5:28 PM

        The way NFL commentators harp on “running, defense and pressure on the quarterback as though they were handed down from heaven” is the same as baseball commentators harping on small ball and playing the game the “right way” despite it being shown otherwise.

        “The NFL community (coaches and ownership/organizational structure) at large is extremely risk-averse and poor at incorporating new ideas into the game.

        Many of the same things used to be said about baseball ownership. Now every team incorporates these ideas into their thinking – some better than others, of course.”

        What “new ideas” are every MLB team using? And which ones aren’t NFL teams using? I’d love to see your source.

      • poreef - Feb 1, 2011 at 6:05 PM

        Well, saying NFL commentators do X, is not the same as saying MLB commentators do not do X. I find it equally annoying in each case, but I do think the NFL is more guilty of it as I feel greater deference is given to former players in NFL pre-game analysis shows.

        For baseball, “new ideas” are the incorporation of modern statistical evaluation of talent into the whole organizational structure. A big part of the NY media going nuts over Sandy Alderson was that he brought that mindset to the Mets, a franchise that only a year ago blew a big load of money on an aging Jason Bay, a move I doubt SA would have done and the Red Sox, with Bill James, didn’t do. Modern stats are mainstream in baseball and have brought success alongside them. Of course, some do it better than others.

        For football, what major scheme innovations have we seen in the last 10-15 years? How many coaches would be willing to take the heat on a surprise onside kick, even when it worked in a Super Bowl? To me the only major innovators of this era are the Patriots, from the coaches to ownership. They’ve done an extremely good job of identifying older players that can be signed for less, yet make key contributions to the team. Likewise, identifying late round players who can make quick contributions. Cutting bait on popular veterans who are on the wrong side of their late 20s. They’re playing moneyball. I have a hard time believing that they could do this without some kind of modern scheme for predicting performance.

        But I don’t really see this model being emulated. The key point is that a lot of this is tied up in labor market issues. You can re-work a baseball team much more quickly than an NFL team because the NFL treats its players like moderately-paid-slaves. It is quite rare for players to switch teams if their original team wants to keep them. Get ‘em while they’re young, use them up, spit ‘em out, and then hopefully your new batch works out as well. If the labor market weren’t so regulated in favor of NFL ownership, I think we’d see a lot more NFL innovation in the Patriots mold.

        PS. Screw the Patriots.

      • frug - Feb 1, 2011 at 9:42 PM

        For football, what major scheme innovations have we seen in the last 10-15 years?

        Zone blocking, the Tampa-2, the zone blitz and the WR/TE hybrid (the Dallas Clark types) have come about since 1995. Moreover, many teams have begun using shotgun and 3 WRs as their base sets. (The multi-receiver base set hasn’t been seen since the Run ‘N Shoot experiment of the early ’90s and the shotgun base set is unprecedented at the NFL level). The no huddle offense has also made re-emerged for the first time since the K-Gun Bills and there has been a shift back to the committee back approach to RB usage after the domination of the workhorse back approach that from the late ’90s through the middle of the decade.

      • poreef - Feb 2, 2011 at 1:43 AM

        I think we’re talking past one another; I shouldn’t have said “scheme” in my question and I fully concede that a new thing or two has occurred on-the-field in the past 15 years. What I’m really after is whether the NFL incorporates advanced stats in its personnel decisions. It seems the Patsies do. I’m pretty sure Al Davis died 15 years ago so he probably doesn’t. For the other teams, I feel that the draconian NFL labor agreement restricts innovation because there’s only so much turnover you can make happen even over a few years.

        Improved statistical understanding has almost certainly led to the emphasis on passing and corresponding trend toward situational backs and tight ends who can catch better than they block. Otherwise, I’m dubious that NFL coaches are at a point where they are willing to go against their preconceived notions of what constitutes a good scheme or situation-appropriate play call. Just look at how many coaches employ schemes without the matching personnel. e.g. Dick Jauron employing Tampa-2 with the Bills. This suggests they don’t understand basic statistics, much less the modern ones you would use.

        NFL teams would do well to hire more people like the football defenders on this forum. Enjoy the Super Bowl.

  9. oldhighs - Feb 1, 2011 at 1:59 PM

    The thing is, in football there really isn’t much you can prove using statistics. As someone above pointed out, it’s way more of a team game.

    In baseball, you can factor in things like OBP, BA from the left and right sides of the plate (not to mention how the particular stadium affects those batting averages), and so on. These things frequently have an actual affect on the game.

    • poreef - Feb 1, 2011 at 2:24 PM

      That’s simply not true. It may be harder to prove aspects of individual performance, but you absolutely can prove team-level choices.

      For example, it has been clearly shown that the old canard about teams that run the ball (and play defense) win championships gets it all backwards. Teams that are winning run the ball to waste time in the 2nd half. Passing plays have – consistently throughout the years – higher yards per attempt. This is one of the few things NFL coaches have figured out – the days of the running back shouldering half the offense are basically over. Now they’re there to keep the defense “honest” and to run out the clock.

      • spudchukar - Feb 1, 2011 at 4:22 PM

        Some of us have known this for a long time, it didn’t take any magical statistics to enlighten us.

    • frug - Feb 1, 2011 at 4:56 PM

      Wrong wrong wrong. There is plenty you can prove using statistics especially at the team level. While the team-centric nature of football makes it harder to isolate individual performance it is quite possible to break down the statistical splits of players when it comes to how a player, for example, performs in different downs and distances, number of pass rushers and side of the field.

      • oldhighs - Feb 1, 2011 at 10:36 PM

        All of those statistics are completely dependent on too many variables to be relied upon as proof of anything.

        Hell, if the Seahawks-Saints game this season proved anything it’s that any given Sunday, any team can beat another team. Games like that happen every single week of the season. And a 16-game season is definitely not a large enough sample to take any sort of useful statistical analysis from.

      • frug - Feb 1, 2011 at 11:33 PM

        I guess it depends on what you are studying. While 16 games is a limited sample size, the average starting QB is going to attempt 400-500+ attempts. Throw in sacks, aborted snaps and rushing attempts and the average QB is going to be responsible for 600 or so plays which isn’t all the different from the number of PAs a major league starter will see in a year.

        Now it is unquestionably true that statistical football analysis is always going to be more difficult and (generally) less accurate than baseball, but to say that you can’t prove much is a step too far.

  10. sdelmonte - Feb 1, 2011 at 2:05 PM

    I read more than my fair share of football coverage, and I think that in general there is not nearly the same level of sabermatical analysis. I don’t think I’ve seen anything along the lines of WAR for quarterbacks, let alone linemen.

    • nps6724 - Feb 1, 2011 at 2:16 PM

      You can see team WPA for every game since 2000 here: http://wp.advancednflstats.com/nflarchive.php

      You can see WPA for QBs, RBs, WRs, and TEs since 2000 here: http://wp.advancednflstats.com/playerstats.php

    • frug - Feb 1, 2011 at 4:51 PM

      Check out footballoutsiders.com. While they don’t have a win based stat like WAR, they have a yardage based stat called DYAR which is essentially a football version of VORP for QBs, RBs, WRs and TEs. They also have advanced defensive stats but they aren’t live updated like DYAR since they require analysis of actual game footage. They also have offensive line numbers, but they’re for collective units not individual players (outside some basic stuff like penalties and blown blocks).

  11. ditto65 - Feb 1, 2011 at 2:11 PM

    What is “football”?

    • sdelmonte - Feb 1, 2011 at 2:16 PM

      It’s that very popular game they play in Europe and Africa and South America.

      • ditto65 - Feb 1, 2011 at 6:48 PM

        Oh – I am familiar. Go Red Devils!

  12. umrguy42 - Feb 1, 2011 at 2:13 PM

    Others have said some of what I would mention – the weather and turf are about it in terms of park “context”. (Although personally, I don’t get the “Rodgers is good in a dome, Roethlisberger is pretty well even indoors/outdoors; therefore, advantage Rodgers” argument. No, please don’t try to explain it, I just call BS.)

    I think, however, the “context” most people tend to talk about football players in is their team as a whole. For example, a recent post over at PFT asks if Big Ben could be considered as good as Tom Brady, especially if he wins his third Super Bowl this week. Arguments have gone both ways, but among them is the contention that Roethlisberger wouldn’t be anywhere without the Steelers defense.

  13. frankvzappa - Feb 1, 2011 at 2:24 PM

    i have occasionally heard mention of Manning’s advantage because of the dome, but only in passing…i would say you are pretty close to right about this, Calcaterra…you may find an odd article here or there about it, but i have been reading every article on PFT for years and dont recall a single item…Manning is certainly overrated, and Rodgers is clearly on the level with Brady as the best in the game…the study of baseball is far more advanced, most likely because baseball attracts intelligent people and football attracts your average stupid American, so they have to dumb down the coverage…and for the record, i prefer baseball

    • nps6724 - Feb 1, 2011 at 2:28 PM

      “but i have been reading every article on PFT for years and dont recall a single item”

      That’s because PFT is a rumor and news site. Do you watch the Food Network looking for the weather?

    • poreef - Feb 1, 2011 at 2:34 PM

      I think it’s true that the NFL has not made as much use of modern statistical analysis. I think this goes back to the conservatism of the NFL leadership – they’d rather have a slaves for life than be able to improve their teams by the free market.

      On a related point, NFL owners are almost all 90,000 years old and like feeling like they’re a big part of the show. Do you really think these organizations are going to adapt to use complimicated stats?

      As much as I hate the Patsies, Kraft and Billy Boy have imported whatever players and ideas that will work and built a system around them that changes meaningfully from year to year. Do they win the Super Bowl every year? No, thank God. But they busted offensive records by doing things like going for it on 4th down a lot more than is typical. Last year, the Super Bowl was arguably won (in terms of “momentum”) on an onside kick. That’s another decision based on statistical analysis – surprise onside kicks have a surprisingly high success rate.

      The days when these stats – at least at the level of “which kind of system should we run” and “what plays should we execute in this situation” – are coming.

  14. BC - Feb 1, 2011 at 3:13 PM

    The NFL has changed a lot in 20 years. You used to get the context with folks like Pat Summerall, Curt Gowdy and the like. Now you really don’t. And with rare exception it’s become a pass-happy league, so the focus of every single team is the quarterback.

  15. xmatt0926x - Feb 1, 2011 at 3:59 PM

    Ugghh…Yes Craig..you are correct of course. This sport of football which is America’s obsession and has grown to a multi, multi-billion dollar business and has for years now been Americas pastime is run by morons who never do any analysis. Yes, they are just all dumb and baseball analysts are all smart. Does that make you and Tommy feel better about yourselves? In all seriousness, I love the different elements that each sport brings and I love the fact that we can debate baseballs history and it’s players in a much deeper way than we could ever discuss any other sports history, but do we really need to pick on those big and awful football baddies every chance we get? Geez, do us all a favor and give us the name of the football player who beat your ass regularly in high school. Again, I love both sports so I dont feel any need to defend football over baseball but the insecurity that is constantly shown by baseball people towards football is rediculous and pretty lame. I have to run now so I can see the latest shot towards football by Jason Stark. He’s yet another baseball guy I love to reads but turns into a whiny and wimpy girl when he discusses football.

    • Craig Calcaterra - Feb 1, 2011 at 4:04 PM

      In what possible way is this post picking on football?

      A writer said “football ignores analysis X”

      I said “I find it hard to believe that football ignores analysis X. Can someone tell me if this is really true?

      If you’re seeing football hate here you’re projecting.

      • poreef - Feb 1, 2011 at 5:32 PM

        @Craig

        I wouldn’t say that football commentators ignore context in their analysis. I’m gonna go out on a limb and say none of them have read Jacques Derrida.

        It’s just that their version of incorporating context is at the conceptual level of Joe Morgan – things people know to be true by repetition rather than through analysis.

    • pauleee - Feb 1, 2011 at 5:43 PM

      It appears football fans are as sensitive as Phillies fans.

      //ducks, runs

Leave Comment

You must be logged in to leave a comment. Not a member? Register now!

Featured video

This was 'the perfect baseball game'
Top 10 MLB Player Searches
  1. S. Kazmir (4904)
  2. K. Uehara (4036)
  3. T. Wood (3618)
  4. G. Springer (3305)
  5. J. Kubel (3245)
  1. M. Machado (2967)
  2. T. Walker (2923)
  3. H. Rondon (2886)
  4. D. Pedroia (2876)
  5. I. Nova (2841)