Mad Scientist or Emerging Genius?As far as crazy off the wall projects go, deciding to single-handedly watch every play of every single NFL game ranks up there. In terms of commitment alone you're talking 256 regular season games, which at three hours plus per game is a boatload of hours in front of a tv.Yet this is what KC Joyner has not only set out to do but has apparently accomplished for several years now. Not enough? A wimpy effort compared to your fantasy football exploits? Consider that he is also writing for ESPN during the season which means he can't just sit back and watch say one game a day, 5 days a week throughout the year. He's watching sixteen games in a week! At TwoMinuteWarning.com we were charting games beginning in 1997, but that was with a number of different people involved and at first all we were doing was tracking a few things on drives. So yes there's respect in these quarters for the effort alone. His first book in 2005 drew a lot of attention, including a notable review by Dr. Z of Sports Illustrated. It focused solely on passing plays and had lots of new metrics like QB bad decisions and WR stats by passing route tree. He also had the defensive coverage side covered as well, with stats on how the DB's did when they were covering the intended receiver. A good start no doubt, maybe even excellent out the gate. But he obviously needed to address the running plays too, which is where this year's edition picks up. In his intro he mentions that he is also looking to move to more objective ratings, and an attempt to standardize ratings. His big motto is "quantify everything and then put it in perspective." Thus we get a vertiable mountain of new stats in this year's book covering blocking, success rates, yards per attempt by play type, passes faced on defense, and more. In fact, the book consists of 8-10 pages of stats and some commentary for each team, followed by 400 pages (FOUR HUNDRED!) of pure stat tables. Holy Moley! Now part of the huge heft of the book comes from the fact KC has put a table in the book for each stat sorted by leader, but any way you slice it, that's a whole lot of numbers. The dilemma is what to make of the numbers. The league wide sorts give some context, but it's still hard to know what to think. For an overall rating Joyner puts the SYPA as about the top of his ordering, which from my understanding amounts to "Success Rate multiplied by yards per attempt." What the heck does that really mean though? Here are some of the leaderboards then for SYPA at WIDE RECEIVER for instance:
You can however pick out all kinds of different ways to evaluate the players. With WR it could be a function of the yards per attempt and success, balanced with drops (shown for both accurate and inaccurate passes), interceptions, strips, lost yardage due to drops, and so on. Joyner also notes the defenders on plays, so adjustments for quality of coverage could also be included...the mind races! But really the most exciting part of Scientific Football 2006 for me is the blocking stats since that's an area that has been so lacking in ANY kind of information for individual players. [You have to ask though, how can KC even tell which players are on the field in some cases? The TV broadcast footage is iffy on many shots it seems.]
Again the SYPA is just one metric, you could go with straight blocking success % (which for a good player seems to be in the 80%+ range, with the superstar blockers hitting the low 90's), successful blocks, gap stuffs, pushed into the backfield, allowed backfield penetration, stringouts, pull blocks, 2nd level blocks, 2nd level success %, penalties called against, penalties drawn... So on the downside if you buy this book you will be in some cases overwhelmed by information. The feel of the book is still a little raw, as it probably was with the initial Bill James' baseball efforts. There is no doubt that KC is collecting just a ton of information, but some streamlining and then better overall rating schemes would seem well worth attending to. The other strange thing is that in spite of so much data, there's still chunks unexplored: no pass blocking ratings by player which would be great in conjunction with the run stuff. No running back detail anywhere near the level of the WR/QB stats. No info on missed tackles/broken tackles (I would love to see a tackle% for defenders). No special teams stats. Now it's obvious we should cut KC some slack since he's just one guy, but that's the ultimate issue: what he has done is wonderful but would it not be more efficient to get a few other people involved and share the workload but thereby pick up some of the missing pieces? Only time will tell what KC delivers to us in the 2007 edition. When he branches off into commentary he is not without strong opinions: he calls the Cincinnati Bengals the "Next Dynasty" and questions whether Bills QB Losman's has what it takes in "leadership skills" to be an effective NFL Quarterback. Pointed comments are certainly more fun to read than bland generalities, even if you don't agree with everything he says!
Amazing stats, but can they predict the future?So what we have is a book with very unique, exclusive and detailed stats on NFL players. Great, fantastic, and interesting reading if you are a pro football fan. The audience here at TMW of course has certain inclinations towards trying to predict the future. Can the book help us there? Perhaps.If you're wondering how certain player acquisitions may pan out, this could be a good tipoff. A new linemen comes in, is he going to help the run blocking? Well compare his stats to the guy he's replacing. Likewise if there is an injury during the year and a starter goes down, the book can help give a view on how the backup may fare. Quarterbacks have arguably the most stats and KC's numbers may be telling for how the young guys can develop or fit in with a new offensive coordinator's game plan. Was last year's passing defense a problem of the defensive backs? Again this book can put the spotlight on individual players where before we had no stats whatsoever, zilch, nada. What is disappointing perhaps is that KC does not make more predictions. If he were to say project a won-lost record for each team in the coming season that would for one give a gauge of the predictive power of his data and two, provide some more mainstream friendly chapters. There's only so many of us football stat-heads out there! For fantasy players KC has a separate publication that goes into his fantasy forecast (his Fantasy Football Draft Guide 2006), and you'd have to think that having seen every play of every game KC should have a read that is beyond almost any public writer. Still it remains to be seen whether this knowledge can translate into fantasy success. After all fantasy football isn't graded on success rates and yards per play, but total yards and touchdowns. Having this data at hand in-season could be very valuable for trying to judge upcoming games. Once again KC aims to please with his Saturday Night Notes reports, where he identifies the best matchups each week.
Our recommendation?
Overall KC Joyner's work is clearly innovative and this kind of undertaking is desperately needed to fill in the oh-so-many-holes in the official NFL statistics. While this book has a lot of meaty material for readers to dive into, there's still the sense that it's incomplete. Once KC adds in pass blocking stats, tackles/tacklebreaking, yards after the catch, yards after contact, running back "bad decisions" and other ideas that must surely be percolating in his head, then we will have the robust, encompassing data set we all want.
One idea that springs to mind is could the FootballOutsiders charting be merged with KC's to provide an even better picture of the NFL? With the FO gang tracking formations, pass-rushers sent and other neat details it might work! As it is, we can heartily recommend you buy this book and we are already looking forward to Scientific Football 2007.
|