Expected Failed Completions is another step in using game charting to break down the air and YAC components in a successful connection. We look at a decade of results and hone in on 2015.
16 Oct 2005
Time for another look at the Football Outsiders mailbag. Most of these questions came via the contact form which you can use to e-mail any of the writers. This mailbag is not just for Aaron, but for any question you might have about any FO article. While we will answer questions posed in discussion threads, we are much more likely to answer a question asked through the contact form. When those discussions get filled up with comments, they get hard to follow.
Be aware that we reference plenty of our innovative FO stats here, not to mention their unfamiliar terminology, so if you are a recent addition to the readership you might want to read this first.
Jersey: One thing I'm confused about [from the DVOA ratings] is this line: â€œFOX RANK represents the FOXSports.com Power Ratings which are 67% 2005 DVOA, 10% 2005 pre-season projection, and 23% a special weighted DVOA for 2004 that includes the playoffs.â€?
Is this something FOX forced upon you in the agreement? Or is it something that was always included but is now sponsored by FOX or something?
Aaron Schatz: No, the "FOX version" power ratings were my idea. Early in the season, we're going to end up with some results that look very strange. We've got small sample size and we don't know yet which teams have really improved and declined, and which teams just had one or two bad games. It was always my intention to develop an "early season weighted DVOA" that would use the previous season's numbers and the preseason projections to give us a better idea of how to rank teams early in the year. Unfortunately, because of the issues with my father, I was unable to scientifically create this formula before the season began. So the "FOX version" formula is cobbled together.
I don't know if the "FOX ratings" make it easier to introduce DVOA to the masses, or harder. On one hand, some teams are ranked closer to how we really think they are, and closer to how other websites have them: New England ninth rather than 19th, for example. In other cases, we look silly because teams that were very good at the end of 2004 have started badly (Buffalo). The initial feedback seems to be that people judge our ratings based on what we say about their favorite team. San Diego fans think DVOA is keen. Denver fans think I'm the dumbest human being on earth. I like the Denver fan message boards where people say that our ratings are crap because we're still counting 2004 performance as part of the formula. Um, actually, if you look closely, you'll notice that Denver would actually rank lower if we only counted the first five games of 2005.
My biggest challenge right now is getting the general FOXSports.com readership to understand what our power rankings are trying to accomplish and why they look so different from other rankings on the Internet. FOX is hugely supportive of FO and nothing is watered down, but I know that the FOXSports.com editors understand us more than the public does.
One more comment about power rankings. Doing these rankings for a major website has really made me notice how totally arbitrary everyone else's rankings are. We all agree that Dr. Z is one of the most knowledgeable football writers in America, right? But he has the same "what have you done for me lately" problem as everyone else. After three games, he had the Patriots listed at #2. The next week, after they lost to San Diego, he dropped them to #11. This week, after they beat Atlanta, they were back to #3. That's crazy. Were last week's 2-2 Patriots that much different from both this week's 3-2 Patriots and the 2-1 Patriots of two weeks ago? Did anyone outside of Washington really think that the 3-0 team that snuck past Chicago, Dallas, and Seattle was better than the two-time defending Super Bowl champions, just because the Pats had lost to San Diego?
While Dr. Z had the Patriots going from 2 to 11 to 3, the FOXSports.com power rankings had them going from 6 to 8 to 9. That just seems more reasonable to me.
Ruben Duran: This occurred to me when I was watching a game last week, and the
dumbasses commentators were talking about draft/contract holdouts: is it possible to NOT enter the draft, and just wait to get picked up as an undrafted rookie, a year or so later? I know it would limit the player's bargaining power, especially if that player was talented, but wouldn't it also allow him to go to whichever team he wants, eliminating the Elway/Eli tantrum? Also, I remember the Texans drafted Drew Henson, even though he had committed to baseball, so I'm thinking of something a bit more complex and underhanded on the part of the player.
Specific example: Player A has a standout season with his college. He is expected to go in the top 5 of the draft, but Player A knows he wants to play for one specific team/market (for whatever reason). Strategically, he decides against hiring an agent, and tells the media he'll be returning for a fifth year at THE University of Michigan (where all awesome players go). He waits until the day after the supplemental draft, and then announces his intent to try out at training camp for his favorite team, and forgo his senior year.
Tim Gerheim (resident ex-NFLPA intern): My instinct is "no" But I can't find anything that would definitely preclude that. The rules that govern the draft are in the NFL's bylaws, not in the CBA. (That, incidentally, is what allowed the Maurice Clarett lawsuit to get off the ground; the fact that they're incorporated by reference into the CBA brought it crashing back down.) The bylaws aren't so easy to get your hands on as is the CBA. But I was able to find some information on the NCAA's website.
As a general rule of thumb, an individual can petition the NFL to declare for the draft three years subsequent to his/her high school graduation date. Four years subsequent to high school graduation, an individual is immediately eligible for the draft.
It's certainly possible for a player to be four years removed from high school graduation and yet have at least one year of college eligibility remaining. In fact, I think it happens all the time. So players who are draft-eligible by virtue of their age go undrafted on a regular basis, which I would think (and the CBA says) makes them free agents.
As a practical matter, however, I don't think it would come up very often. It has to be a ruse by the player; if the NFL's teams think he's going pro, someone is likely to draft him, at least in a late round (a la Drew Henson). So to all outward appearances, the player has to be headed back to college. I'm pretty sure in order to play college football, a player has to still be taking classes, even if it's just a ballroom dancing class, so I'm not sure that players who are actually going to graduate would be able to pretend that they're planning to continue to be college athletes. (It is possible to play college athletics as a graduate student.)
So the first reason that Eli didn't pull your trick is that everyone and their dog knew that he was going pro after the 2003 college season. He might not have had any more college eligibility left, in which case it would have been an empty threat. But even assuming he did, he couldn't risk holding out of the draft. Some team somewhere along the line would have suspected that he was going to go pro, and it would have drafted him, maybe in the fifth round. Well, then the team drafting him is only going to allot fifth round money (maybe a little extra) to him, because it can only push their rookie salary cap pool, which is based on the specific draft slots it has, so far.
The NFL salary cap system expects the top pick in the draft to get an $8 kajillion signing bonus, but a fifth rounder only gets probably $300k to sign. If he succeeds in not getting drafted as you suggest, it's actually even worse for him financially, because guess how much the NFL allocates to teams for each undrafted free agent? Exactly $0. He'd be lucky to get a $15,000 signing bonus because there's just no money left for him. Granted, he could sign a one-year contract and demand a big contract as a nominal free agent the next year, but as an exclusive rights free agent he would have no bargaining power whatsoever, so he would only get a goodwill contract (see Gates, Antonio) rather than a "crap, he's gonna leave!" contract (see Shockey, Jeremy).
So I think the long and the short of it is that it sounds like a nifty loophole, but in point of fact the salary cap system makes it severely unappealing.
Bill Moore: In regards to Drew Henson, the Houston Texans owned his rights when they traded those rights to Dallas. He was drafted in the sixth round of the 2003 draft (despite leaving Michigan in 2001), and Houston held those rights for one year. If Houston was unable to sign him or trade those rights by the 2004 draft, he would return to the draft and become eligible for another team to draft him. Had Houston not traded his signing rights, and had no other team drafted him [that wouldn't have happened given the number of teams interested in him, but work with me], he would have been eligible to sign with anyone as an undrafted free agent.
As for the analogy, one can not enter the NFL without making himself eligible for the NFL draft. As a result, you can not become an undrafted free agent if no team has ever had the ability to draft you. NFL Draft eligibility comes in two forms. To become automatically eligible for the NFL Draft, you must fulfill your college football eligibility. Otherwise, you must seek entrance to the draft. As a result, Player A, who announced that he will return to Michigan, and then chooses not to, would have to wait to declare himself eligible for NEXT year's draft since he has not completed his college football eligibility.
[Side note: this is probably the reason I have never been drafted given that I did not fulfill my college eligibility at Football power house Babson.]
Player A would have to fulfill his college time, and then announce that he intended to go into landscaping and HOPE that no spends a draft pick on him. Given his talent level, the likelihood of him going undrafted is quite low.
Interestingly, there is the clause in the CBA that says the following:
"Return to College: If any college football player who becomes eligible for the Draft prior to exhausting his college football eligibility through participation is drafted by an NFL Club, and returns to college, the drafting Club's exclusive right to negotiate and sign a Player Contract with such player shall continue through the date of the Draft that follows the last season in which the player was eligible to participate in college football, and thereafter the player shall be treated and the Club shall have such exclusive rights as if he were drafted in such Draft by such Club (or assignee Club)."
As I read that, it says to me that if you really don't want to play for the 49ers or Saints or whomever, you can declare yourself eligible a year early, announce your intention to go back to school, and hope a better team that doesn't need an impact player today drafts you with a later pick. You go back to school, and sign with that team the following year. That strategy will never happen for four reasons, 1. MONEY, 2. MONEY, 3. MONEY, and 4. the 49ers may draft you anyway in a later round.
Keith Hudson: I saw Titan Bo Scaife listed on the "Wide Receivers 2005" stats page here. However, all the sources I could find (titansonline.com, nfl.com, even Wikipedia) show him as a tight end. Is this because he's generally used as a third tight end? Is it an oversight? Is it something else that I can't think of?
Aaron Schatz: It's me screwing up. I have to mark each player's position and I have to remember to go make sure whether guys in the 80s are WR or TE (and whether guys in the 40s are RB or TE). I must have marked him WR without checking him. I'll fix that and he'll be listed among the TE next week. Football Outsiders: Not Afraid to Admit We're Wrong!
Israel Pikholz: How exactly are supplemental draft picks determined? I know it has to do with losses by free agency, but how much does success come into play? For instance, should the Steelers be rooting for Plaxico Burress to go to the Pro-Bowl?
Michael David Smith: I believe you are referring to the compensatory draft picks that teams can receive, not the "supplemental draft" that takes place in June and brought us Tony Hollings and Manuel Wright. Compensatory draft picks go to teams that lost more unrestricted free agents than they signed. So if your team loses six unrestricted free agents and only signs two, your team will get four compensatory draft picks. The picks are awarded at the end of either the third, fourth, fifth, sixth, or seventh rounds, and the determination of how high a pick a team gets is based on the size of the contracts of players lost. So a team that loses big-money players would likely get third-round compensation, whereas a team that loses only budget-priced free agents would likely get seventh-round compensation. (For an example, I found this story discussing the rewarding of 2003 compensatory choices.) The Giants don't need to root for Plaxico Burress to do well because the compensation is awarded based on contract size, not on-field production.
Tim Gerheim: I'm not sure it's that simple. As I recall, Len Pasquarelli and/or John Clayton often refers to the compensatory pick formula that the NFL uses as one of the most arcane and closely-guarded formulas in sports. I found this description on a rec.sport.football.pro faq, which is obviously not the most authoritative source:
Compensatory free agents are determined by a formula, developed by the NFL Management Council, based on salary and performance. Not every free agent lost or signed by a club is covered by this formula. The number of picks a team receives equals the net loss of compensatory free agents up to a maximum of four. In some cases team receive picks even though they did not suffer a net loss of compensatory free agents. Under the formula, the compensatory free agents these teams lost were ranked higher than the ones they signed (based on salary and performance). No compensatory picks are higher than the third round.
That does indicate that performance is a factor. I think the short answer is that I don't really know, and the NFL keeps that information close to the vest.
(Name Withheld): You are extreamly under qualified to select powerankings. I hope you have no offspring.
Aaron Schatz: That was the Football Outsiders Complete Non-Sequitur Hate Mail of the Week. Oh, and you lose, sucker.
Buzz Newberry: Aaron, in this week's DVOA commentary you mention your rankings vs. number one, two, and three receivers. Is there somewhere I can see that?
Aaron Schatz: You aren't the first person to ask about this. I've put together a table and posted it on the team defense page. Just be aware that this table isn't yet built into my automatic table builder program, so it may not be updated every week. (The table builder programmer, John Argentiero, is currently in a busy period of his real job.)
Dan: Have you considered reporting separate "Offense Variance" and "Defense Variance" ratings, in addition to the Variance rating based on total DVOA? A team could have low Variance just because it has happened to play better defense in the games when it was worse offensively, but that doesn't mean that it's played consistently. That's one reason, I think, why Washington's Variance is so absurdly low (1.7%) -- their offense was better than before against Denver, but their defense was worse. You might even break it down into rushing and passing Variance for both offense and defense.
P.S. I really think that Estimated Winning Percentage would be an easier statistic to interpret than your modified version of Estimated Wins.
Aaron Schatz: Hey. Yes, in the book we did separate offensive and defensive variance ratings. Here are the numbers through Week 5. Teams are ranked from least consistent (#1, highest variance) to most consistent (#32, smallest variance) and are in order by total variance.
There's no doubt that total variance doesn't give the whole picture. Buffalo and Cincinnati, for example, have been fairly consistent on offense, but up and down on defense. (Of course, Cincinnati has been consistently good on offense, and Buffalo has been consistently bad.) Houston, San Diego, and Jacksonville have been the opposite: consistent on defense, inconsistent on offense.
San Diego is particularly interesting because the overall DVOA doesn't have a very high variance but the offensive DVOA does -- the Chargers had two huge games against the Giants and Patriots, but rate as close to average in their other three games. Because the defense comes out near average in all five games, the Chargers don't have a game that comes out as particularly negative overall. Most teams, even after five games, have at least one big positive game and one big negative game, which is why the "total" variance numbers are so much higher than the offensive and defensive variance numbers.
However, it turns out that you are wrong about Washington: both the offense and defense have been equally consistent this season, although the Denver game was the best offensive performance:
|WAS DVOA||Week 1 vs. CHI||Week 2 at DAL||Week 4 vs. SEA||Week 5 at DEN|
I'll try to remember to add these to the team offense and team defense pages in the future. (As I said above, they can't be added into the automatic table builder right away.) As for Estimated Wins, it is much easier for football fans to think in terms of total wins. What is Philadelphia's record right now? Do you think "3-2", or do you think ".600"? Most people think 3-2. That's why I list Estimated Wins as a total, not a percentage.
Sid: Has the PFP cover curse been mentioned in an article or anything yet?
The only one not affected was Matt Hasselbeck, and that's because his name was misspelled.
Aaron Schatz: Wouldn't the curse only work if the two players we said negative things about on the cover (Martin and Vick) were actually having really good years?
Curious: Any thoughts on why Carolina is so far down the DVOA in 2003? They went 11-5, and went through DAL, STL, and PHI on the way to the superbowl, yet they have a negative rating and are ranked 20th(!)
Aaron Schatz: Ah, you must be new. Yes, Carolina drove me nuts during our first season, and that's covered in the archives. This is probably the best article to read about the strange case of the 2003 Panthers.
Next week: We hope to finally get to questions from the 1998 and 1999 ratings. Really.
35 comments, Last at 20 Oct 2005, 10:30am by B