CSS Button No Image Css3Menu.com

Baseball Prospectus home
Click here to log in Click here to subscribe

Chat: Aaron Schatz

Chat Home

Welcome to Baseball Prospectus' Tuesday November 08, 2005 1:00 PM ET chat session with Aaron Schatz.


Aaron Schatz, the founder of FootballOutsiders.com, is the lead author for Pro Football Prospectus 2005.

Aaron Schatz: Hello, folks. To introduce myself briefly, my name is Aaron Schatz, I run the NFL stat analysis website called FootballOutsiders.com and write the weekly power rankings for FOXSports.com. I also am the head writer of Pro Football Prospectus 2005, still available from your favorite bookstore or online book seller, and the upcoming Pro Football Prospectus 2006, which is like PFP 2005 except it's one larger. Thanks again to BP for partnering with us and for allowing me to steal space for a chat now that the baseball season is over. I've got Super Furry Animals as today's soundtrack and I'm ready to take your questions about the current season, fantasy football, or the role of statistical analysis in the NFL.

Todd (MD): Aaron I've seen a site that promotes the use of history to rate performance...does DVOA take into account past performance by teams when rating them?

Aaron Schatz: Before answering this question, since this week's latest DVOA ratings are not yet up -- that's Defense-adjusted Value Over Average, our main metric for measuring teams -- I'll share the top ten for you chatters. These are the "weighted" ratings, with more strength for more recent games.

1 IND 8-0
2 CIN 7-2
3 SEA 6-2
4 NYG 6-2
5 SD 5-4
6 DEN 6-2
7 JAC 5-3
8 DAL 5-3
9 PIT 6-2
10 WAS 5-3

As for this question, no, DVOA is based on the current season only. For FOX, I did a power rating that included last season's numbers for a while until we had enough games to feel confident about 2005 numbers, but that was sort of cobbled together. Next summer I'll work on something that makes that a bit more scientific, but once you hit midseason it is pretty clear that a number of teams have improved or declined from their past performance.

Evan (Vancouver, BC): Hey, Aaron. Any luck finding the data for those CFL translations, yet?

Aaron Schatz: Nope, I'm not even sure where to find historical CFL stats, although we're considering building a USFL translation system to answer the question of whether Herschel Walker deserves to go into the Hall of Fame.

Kotch (Washington, DC): Aaron: How do you stop Antonio Gates?

Aaron Schatz: Strychnine

jpyuda (Alexandria, Va.): How does the T.O. suspension change the Eagles. Does it make them better? I feel like "team chemistry" is usually only an issue in extreme cases. But if ever there was an extreme case, T.O. is it. Also, I feel like getting rid of him might improve the play calling and bring that closer to balance.

Aaron Schatz: I think that a lot of the BP guys agree with me on this -- the issue with chemistry isn't that it doesn't exist, but rather that it is impossible to measure objectively. So trying to build your team around something which is mostly guesswork may not be the best idea. That being said, there are some clear cases and yes, this is one. But better chemistry isn't going to make Donovan McNabb healthy. It isn't going to make Dirk Johnson healthy. And it isn't going to make the Giants, Cowboys, or even the New Potatoes into losing teams. As for play calling, I don't think the lack of running plays is an issue with T.O. The Eagles need an inside runner to pair with Westbrook, and Andy Reid has got to start calling some runs so that opposing coaches have to formulate an entire defensive game plan instead of just half of one.

Greg (Boulder, CO): Football Prospectus is a great addition to the BP family. I noticed that quarterbacks' DPAR is generally 2-3 times higher than that of running backs. Does this reflect the relative value of the positions, or is it more a function of their respective replacement levels?

Aaron Schatz: Thanks for diggin' it. There's two issues here. One is that quarterbacks are involved in more plays than running backs. The other is that the range of performance for passing plays is much wider than for running plays. After all, a good pass is 10+ yards, a bad pass is zero. A good run is usually around 5-10 yards, a bad run usually gets a couple yards anyway.

Fantasyowner (yahoo): Didnt you say Kevin and Julis Jones would lead the league in rushing? I traded down from the #2 spot (where I could have had Tomlinson) and drafted them. Reuben Droughns and marion Barber are my highes scring RB's. Thanks!

Aaron Schatz: Oh, yes, this. I'm definitely wishing at this point that we had stuck Willis McGahee on the cover instead of Kevin Jones. We've gotten a ton of complaints about the Jones projection, which is in many ways a case of the broken tile syndrome. Look up at a new ceiling, and if one tile is broken, that's the one you'll notice, even if the rest is perfect. Many of our projections were accurate, but the mistakes are more obvious. (We did a lot better with our predictions of who would decline -- Tony Gonzalez, Jamal Lewis, Manning and Culpepper -- than our predictions for who would improve.)

That being said, we did mess it up, and when we improve the system next year we have to ask why. With Kevin Jones and Domanick Davis I think the problem is that we didn't do enough to take into account the effect of a bad offensive line. Julius Jones was impossible to predict -- injuries to guys in their early 20s are just random.

Oh, and nowhere in our book did it tell anyone to trade out of a slot where they could draft LaDainian Tomlinson. He was projected with more touchdowns and more receiving yards than the Jones brothers, and was on the top of the fantasy value list in the appendix, with Shaun Alexander right behind him. I'm not sure what you were thinking there.

lpolinsky (Philadelphia): Indy going for 2 after making it 40-21 -- legit football move, or bush-league rubbing-it-in?

Aaron Schatz: Legit. What the heck does one point do for them there?

Doug (PA): In PFP 2005 you named Kevin Jones as the future NFL rushing champ...who do you predict now?

Aaron Schatz: Hmmm. Good question. Don't have a midseason stat projection system or anything like that but based on the current numbers and the remaining schedule, I would say Shaun Alexander. Significantly easier schedule than James, McGahee or Tomlinson, and Dunn will lose a few carries when T.J. Duckett comes back. By the way, did anyone see Duckett on the sidelines this week? Man, he's got a great Beard. He's rockin' some serious Isaac Hayes facial hair.

Mac (NJ): Is there any kind of Pathagorean record for NFL teams (what a teams record should be based on its PF and PA) or are there just not enough games for this? Will a team with a .500 or better record and a negative point differential have things "catch up with them" eventually or does this concept not apply to football at all like it does baseball?

Aaron Schatz: Yep. The exponent we use is 2.37, developed by Daryl Morey from STATS Inc. a few years ago, although I've seen 2.85 used elsewhere. Things do catch up -- in particular, teams that significantly underperform their Pythagorean despite a winning record, like the 2002 Chiefs, tend to get a lot better the next year -- but the sample size is going to make it less accurate than baseball. I mean, the Packers right now have scored more points than they have allowed solely because they stomped the New Orleans Saints of San Antonio-Baton Rouge-Oklahoma City.

I wrote about this for the NY Times last year -- out of the 17 Super Bowls played since the 1987 strike season, 11 have been won by the team that led the NFL in pythagorean wins, but only 7 by the team that led the NFL in actual wins. That includes the 2004 Pats, 2000 Ravens, 1999 Rams, and 1997 Broncos.

Patton1941 (NYC): Is there any way to separate a RB's ability/worth/value under DVOA from the quality of his offensive line or the scheme?

Aaron Schatz: Well, we've got the adjusted line yards system that tries to do that -- not scheme, but offensive line. It works differently from DVOA because it is based on yardage, not yardage and first down conversions combined, but you learn some interesting stuff from it. http://www.footballoutsiders.com/stats/ol.php

ssimon (Pelham, NY): Aaron, in Monday's column, Peter King suggested that we should be having a discussion about Tiki Barber as the NFL's "co-best-back" along with LaDanian Tomlinson. Is this a legitimate comparison? Can you give me a baseball analogy Tiki:LT :: X:Y ?

Aaron Schatz: Hmmm. They do have similar numbers over the last couple years -- remember, Tomlinson's numbers suffered last year because of a groin injury. I actually thought about this one. Tiki is to LT as Pedro Martinez is to Dontrelle Willis. Both are great pitchers, but Willis is more durable and more valuable in fantasy solely because of a stat which is in large part determined by his teammates rather than his own performance (wins in baseball, TDs in football).

Joe (Cleveland): Does ESPN know how to cover anything but Brett Favre and Terrell Owens? How do you react to this constant coverage and sesationalism?

Aaron Schatz: You forgot Kobe Bryant, the Red Sox, and the Yankees.

Curtis Martin (New York): Is my career just about over or is my lack of production a product of a bad offense and injuries to the O-line?

Aaron Schatz: I've written about this a few times. Every single running back who has carried the ball at least 370 times in one season has either torn his ACL, lost his effectiveness within two years, or finished his career with the Raiders before hanging up his goggles, shaving his head, and becoming the least-articulate sideline reporter in the history of televised sports. Unless NBC wants to try Curtis in the Eric Dickerson role next year, I think you are seeing result number two. You also have to remember that Martin had the best season of his career at age 30, which is extremely rare, since most RB start to decline after age 28. And part of it is the offensive line. How's that Mike Nugent thing working out for them?

coneway (austin, tx): Would you agree that in a certain way football has been ahead of baseball in its willingness to hire front office execs who weren't former players?

Aaron Schatz: Yes, definitely. Part of this, I'm guessing, was the family-owned nature of most NFL teams for most of the league's history. The owners often ran the teams, and they didn't play, so they had no problem hiring smart people who didn't play in executive roles.

sporer24 (Austin): Would the Lions be better off just punting on 1st down?

Aaron Schatz: Keep it down, the people on the other side of Michael David Smith's office are comparing about the swearing right now.

adamcarralejo (santa barbara): Hey Aaron, love what you guys at FootballOutsiders do. I love watching football, but i cant help but feeling that the fans in LA are getting the short end of the stick from the NFL. We contribute to their TV ratings and revenue, but we will never be able to watch an LA team occomplish anything. Is it the NFL's fault, or something else? Id love to hear your take on the situation.

Aaron Schatz: People often complain that L.A. had two teams and if both moved, this meant that the people of L.A. couldn't support them, so they don't deserve another team. You know, I lived in Orange County for a long time and I never got the sense that L.A. didn't want a team. What happened was that the Rams and Raiders both blackmailed their respective cities (Anaheim and L.A.) the way sports teams do all the time and the Californians said no, we're not spending tons of tax money to build you new stadiums, and Oakland and St. Louis said, hey, come enjoy our tax money, and so both teams left. To this day L.A. refuses to kowtow to the owners who want them to spend public money on a big new football stadium for the NFL. If they move a team there, that team will get plenty of support provided that it doesn't turn into a perennial loser, but the owner of that team is going to have to take care of his own stadium situation.

sporer24 (Austin): Is Priest Holmes done as a top-flight fantasy back?

Aaron Schatz: Yep. Age, knees. RB go quickly. Another interesting note: our player comment for Holmes in PFP 2005 pointed out that his role in the passing game severely declined last year, and this seems to be a sign of decline in a running back. Emmitt Smith, Thurman Thomas, and Marshall Faulk all saw their receiving numbers take a steep drop a year or two before their running effectiveness did the same.

Oh, and I forgot to mention in the previous answer, we never called it "The O.C." What is that nonsense?

Ben (Chapel Hill): Hey Aaron, thanks for chatting. I was wondering if you have any plans in the works to try to apply this stuff to single game decisions and predictions. Also, have you guys gotten any calls for consulting from any teams?

Aaron Schatz: Yeah. Would love to build some sort of system to try to predict single games, although it would need many more variables than just our stat ratings -- injuries being the most important issue. It's on the to do list, but the to do list is absurdly long, because football research is still so young that there are tons of ideas, and I'm the only guy who does this full time, and even I don't have much time to do new research during the season because I'm busy writing the articles that pay the mortgage.

And yes, I've spoken with a number of teams, although not the Patriots. I'm only a half hour away guys, I could drop my daughter off at my mom's and be over in five minutes. I'm in the phone book, Bill, look me up.

Doug (PA): Is there any good reason that DVOA hasn't been applied to college and high school football yet?

Aaron Schatz: High school football?

Greg (Boulder, CO): Any thoughts on performing your analysis tools at the college level?

Aaron Schatz: But seriously, yes. Obviously the sample size issues will be even more annoying with just ten or eleven games and tons of player turnover each year. But I'd be interested to do something. Once again, the issue is a) time and b) inclination. I'm not a college football fan. But a lot of people are, no doubt. Honestly, if anyone is interested in doing college stat research, e-mail me, and I'll put your name away and contact you in February or something.

Steve (Manalapan): When espn average user picks are closer to reality than your stats book, then there is a slight problem. A big selling point of your book was your accurate fantasy projections! Can I have my money back?

Aaron Schatz: Hmm, they are? I'd be curious to see that list. Is that the average user pick from August, or June? Obviously, when the book goes to bed in June, it's going to be wrong about injuries and depth chart changes in July and August. As I said earlier, the basic feeling I get from comparing our projections to the conventional wisdom before the season is that we didn't do very well on predicting which young players would improve and we did much better than conventional wisdom on predicting which older players would decline.

Wayne (Detroit): Why do you have Pittsburgh rated lower than Cincinnati?

Aaron Schatz: Our ratings are based on the idea that measuring individual plays provides a much more accurate sample size than simply measuring eight binary win/loss decisions. So while Pittsburgh's win over Cincinnati was a big boost to the Steelers' rating and dropped the rating for the Bengals, it doesn't cancel the rest of the season. Pittsburgh nearly lost to two bad teams over the last two weeks, Green Bay and Baltimore. The Bengals have a dominating win over a 5-3 division leader on their resume (Chicago, a team nobody wants to think about). Add up all the plays, and the Bengals come out higher. That doesn't mean I think they would win a playoff game between the two teams. But they would if Charlie Batch was starting, that's for sure.

mreedy (Lawrence, KS): I know this isn't quite what you're here for, but you know a lot about how people play fantasy football. I'm in a HUGE fantast league (20 teams and 15 man rosters). Player availability is, to say the least, an issue. What would you say is the ideal # of teams and players for fantasy football?

Aaron Schatz: 12 teams, 8 starting players on offense, and either a defensive team or individual defensive players depending on your inclination.

Jake (Los Angeles): LaDainian Tomlinson's per carry averages are pretty pedestrian (around 4.0-4.1 over the last season and a half). Can you be the "best back in football" with those kinds of numbers?

Aaron Schatz: Discount last year, which he played with a groin injury. This year, part of the low yards/carry (4.3) is the schedule, although I don't think that gets easier in the second half. I don't think he'll ever reach the heights of his 5.3 yards per carry in 2003, but considering that he gains a lot more of his yards on long runs (i.e. without as much help from his offensive line) compared to, say, Edgerrin James, I think it is still safe to consider him the top guy. Tomlinson and Alexander are currently tied atop our DPAR (Defense-adjusted Points Above Replacement) rankings for running backs.

Lightning round!

Undecided (Florida): Who's better in fantasy from now on: Vick or McNair?

Aaron Schatz: Tough choice, probably McNair because Vick still has to play the CAR and TB defenses twice, plus Chicago.

Doug (PA): Only four times in 39 years of Super Bowls has the league's leading rusher played for the Super Bowl champion team -- three of those times it was Emmitt Smith...what would DVOA say about that?

Aaron Schatz: Passing is more important than rushing.

Dennis (Newark): You mention about the book going to bed in June. Shouldn't this be a huge warning on the cover of your book? Depth charts are a lot less predictable in football than other sports. If you physically can't produce a book closer to the start of the season, shouldn't you admit that there are going to be major problems?

Aaron Schatz: It's no different than a fantasy magazine. The Internet is a faster medium than print. We offered updated projections on the website for folks like Mike Anderson and Kyle Orton.

dokomoy (Los Angeles): Is it just me or is mentioning the w-l record of a qb almost as dumb as giving pitchers wins and loses? Leading upto yeasterdays game you couldn't go more than 30 seconds without someone mentioning Peytons career record vs the pats or the rest of the league.

Aaron Schatz: Yep.

shamah (DC): So what do the Jets do for the rest of the season and the offseason? The defense has some young, core guys to build around, but the offense is a mess: they let the wrong RB go and the QB situation is murky. Do they blow up the offense and start over? What about all the money we owe Coles and Pennington?

Aaron Schatz: Call your favorite Tennessee or San Francisco fan and ask them what the last couple years have been like.

mreedy (Lawrence, KS): It looks like the Chiefs need to start over with an eye towards 2007-2008. Besides LJ and Derrick Johnson, do they have any keepers?

Aaron Schatz: Yes, but they're all young defensive players because they keep using all their draft picks on the defense while the offense gets older and older.

ssimon (Pelham, NY): Is the center position really this important? Kevin Mawae and Matt Birk both go down, and so do their respective teams... Are the Jets and Vikings reeling because they lost their C, or is that just one factor in *many*?

Aaron Schatz: Offensive linemen are important. Never talked about, but important. But it is a bigger factor with the Vikings than the Jets, and the loss of one player never destroys a team entirely, there are plenty of other issues.

Cedric Benson (Parts Unknown): Hey, whatever happened to me?

Aaron Schatz: Apparently you forgot to call Philip Rivers so he could tell you that long rookie holdouts are going to mess with your career.

Mike W (Chicago): If Derek Jeter = Brett Favre, then who is Favre's Tim McCarver? Michael Irvin?

Aaron Schatz: Tom Brady is Derek Jeter, not Favre. Favre is, I dunno, Mike Piazza or something.

PJ (Parsippany): A lot of Tiki's fantasy value was lost once it was known that he wouldn't be getting goaline carries. Coughlin (sp) did the same thing in Jax with Fred Taylor. Do you guys pick up on these type of coaching tendancies when evaluating players?

Aaron Schatz: 95% of the time these kinds of preseason announcements turn out to be complete nonsense. Jacobs happens to be part of the other five percent. Fantasy football is always going to have this problem where projections are difficult because we give far more weight to something that is extremely variable (touchdowns) than we do to something that is far more constant (yards).

oskithebear (nyu): How likely is it that the Colts run the table?

Aaron Schatz: at CIN, PIT, at JAC, SD, at SEA. That's five teams in the DVOA top ten. It isn't going to happen.

sporer24 (Austin): Will PHI and NE make the playoffs? One, both, or none?

Aaron Schatz: Patriots will win their division at 10-6. I hate to say this, because I picked them to win the Super Bowl, but I think the Eagles are gonna fall short.

Mike W (Chicago): They threw up a stat last night about red zone efficiency of one of the teams. Is that random from year to year, and do some teams over- or underperform expectations and if so, do you adjust for that in your predictions for next year, much like 3rd down efficiency? BTW, you and FO rock.

Aaron Schatz: I think I answered this in the last FO mailbag, check footballoutsiders.com.

ElAngelo (NY, NY): Are the Redskins for real or a mirage that will come back to earth (much like their 7-1 start a few years back that ended up in them missing the playoffs)?

Aaron Schatz: This is why small sample size is so difficult to deal with -- the Redskins won three games they should have lost, then lost two games they should have won, then had a huge win, then had a huge loss. I have no idea how good they really are. Probably not as good as the Giants or Cowboys.

Tim (Bay Area): By answering a question referencing Jeter with a "Tom Brady is Derek Jeter" line, did you intentionally open the flood gates for the Brady/Manning debate to bleed into this chat? Common, say it: Manning is A-rod.

Aaron Schatz: Trent Dilfer is definitely Omar Vizquel.

Bob the Hobo (under the boardwalk): Does Detroit still have a shot at *hiccup* the NFC North?

Aaron Schatz: OR-LOV-SKY!

Gary (NJ): I'm with you on the Brady/Jeter thing. If Manning and Brady trade places for the past four years, do the Pats and Colts trade championships?

Aaron Schatz: No, no, no, no, no. Rodney Harrison, Richard Seymour, Ty Law, Willie McGinest, Tedy Bruschi, Mike Vrabel, Bill Belichick, Adam Vinatieri -- people think these guys didn't matter because they aren't quarterbacks? Seriously, that's disrespecting Rodney Harrison. I would hide the children in case he finds out where you live.

Aaron Schatz: OK, people may be able to tell by now that I need to eat before I start getting really goofy. Thanks for all the questions, even the negative ones about the fantasy projections, and perhaps we'll be doing this again sometime. See you over at FootballOutsiders.com.

Baseball Prospectus Home  |  Terms of Service  |  Privacy Policy  |  Customer Service  |  Newsletter  |  Masthead  |  Contact Us