I’ve fallen off the wagon in terms of the 2006 stat analysis, but here’s your next installment. So far I’ve looked at success rates for QB’s, RB’s, and WR/TE’s...now it’s time to look at line play.
There is no perfect tool for evaluating an O-line without watching every play of every game and grading each player out like coaches do. I wouldn’t have the time for that even if I had access to tapes of all the games. Just ain’t gonna happen. That said, Football Outsiders use a tool called Line Yards. I’ve used it for my ‘Beyond the Box Score’ bits, and I’m going to dig it back out now to measure run-blocking effectiveness.
Line Yards is once again defined by Football Outsiders as follows:• For a play that resulted in negative yards, the O-line is granted 120% of the effort (i.e. a 3-yard loss would be a 3.6-yard loss for the O-line).
Obviously this tool is far from perfect, as it doesn’t pay heed to the abilities of the people running or throwing the ball. You open up the same-sized hole for Adrian Peterson or Jon Cornish as you do for Paul Mosley or Stevie Hicks, and you’re not going to get the same yards every time. The RB still has to hit the holes and maybe break some tackles. That said, capping the possible gain at least takes an RB’s explosiveness (and the occasional big-play outlier) out of the equation. Whether an RB gains 10 yards or 20 or 80, the O-line did its job, and it gets max credit for it.
• For a play that resulted in a 0-4 yard gain, the O-line is granted 100%.
• For a play that resulted in a 5-10 yard gain, the O-line is granted 50% of the yards over 4 (i.e. an 8-yard gain would be a 6-yard gain for the O-line).
• For a play that resulted in a 10+ yard gain, the O-line get no extra credit—by that point, the runner is into the secondary, and the line won’t get much chance to block. Therefore (if the math in my head is correct), the most credit an O-line can get is 7.5 yards.
So here are the 2006 line yards. Two qualifiers: these are only conference games, and these stats—like most others I use—are only for when the game is within two possessions. If you’re down 30 and the other team brings in its second-string nickel package, you can probably run for 8 yards a pop no matter who you are...and it would prove nothing.
2006 Line Yards (Conference Average: 2.86)
1. Kansas – 222 attempts, 763.5 yards, 3.44 average (Value Over Average* : 1.20)
2. Oklahoma State – 260 attempts, 851.8 yards, 3.28 average (VOA: 1.15)
3. Nebraska – 279 attempts, 877.8 yards, 3.15 average (VOA: 1.10)
4. Texas – 214 attempts, 673.1 yards, 3.15 average (VOA: 1.10)
5. Colorado – 220 attempts, 685.9 yards, 3.12 average (VOA: 1.09)
6. Texas A&M – 330 attempts, 993.8 yards, 3.01 average (VOA: 1.05)
7. Oklahoma – 314 attempts, 878.9 yards, 2.80 average (VOA: 0.98)
8. Missouri – 192 attempts, 532.4 yards, 2.77 average (VOA: 0.97)
9. Kansas State – 188 attempts, 511.4 yards, 2.72 average (VOA: 0.95)
10. Texas Tech – 84 attempts (!!), 217.3 yards, 2.59 average (VOA: 0.91)
11. Iowa State – 174 attempts, 438.5 yards, 2.52 average (VOA: 0.88)
12. Baylor – 104 attempts, 188.0 yards, 1.81 average (VOA: 0.63)
* Value Over Average just signifies a team’s performance/value compared to the conference average. 1.00 = conference average. 1.20 = 20% over the conference average. 0.80 = 20% below the conference average. Et cetera.
Thoughts:• Jon Cornish led the league in rushing, so it would make sense that Kansas would be #1.
Okay, so that works pretty decently for a line’s performance in the rushing category. What about passing? What’s an O-line’s main job in pass blocking? Not getting the QB sacked, of course. It would make sense, then, that a statistic like Sack Rate (sacks divided by passing attempts) could be used. Now, this comes with the same “everybody’s different” issues as Line Yards. A sack rate could depend just as much on a QB’s ability to avoid oncoming rushers and make a quick decision or the overall ability of the WR’s and TE’s to actually get open. But as with Line Yards, the Sack Rate will tell something, right?
• Oklahoma State at #2 surprised me simply because I figured a lot of their yards came off of big runs (I suspected the same out of K-State and was correct on that one). Instead this shows a pretty decent consistency.
• Texas A&M and OU were rated #1 and #2 in most unit rankings coming into this season, but their totals weren’t nearly as high as I would have thought.
• I’m thinking Colorado didn’t run the ball enough, eh? With Bernard Jackson’s arm and the Buffs’ cruddy receivers, Coach Nick Nolte’s offensive schemes didn’t really match the talent. And by “didn’t really match” I mean “was terribly, disastrously at odds with.”
• When games were within two possessions, Texas Tech averaged about 10.5 rushes per game. That’s insane. I mean, I knew it would be a small number—duh, they’re Texas Tech—but that’s...well, insane.
As we delve into Sack Rate, we should take into account the fact that there’s a higher rate of sacks on third (and fourth) downs than on first and second. Makes sense. Third downs see more pass situations and therefore see more blitzes. So since preventing sacks on first/second downs and third/fourth downs are really sort of two different skills, what we’re going to do here is look at both the sack rates for both and hold them as separate entities, comparing them both to the overall conference averages, and figuring out a way to make some judgements based off of that.
First/Second Down Sack Rate (Conference Average: 5.46%)
1. Texas Tech – 1.94%
2. Oklahoma State – 3.94%
3. Oklahoma – 4.13%
4. Missouri – 4.55%
5. Kansas – 4.88%
6. Texas A&M – 4.90%
7. Texas – 5.77%
8. Baylor – 6.15%
9. Kansas State – 6.75%
10. Iowa State – 7.94%
11. Nebraska – 8.24%
12. Colorado – 9.47%
Third/Fourth Down Sack Rate (Conference Average: 8.10%)
1. Oklahoma – 5.00%
2. Missouri – 5.19%
3. Texas Tech – 5.56%
4. Oklahoma State – 6.45%
5. Colorado – 7.55%
6. Texas A&M – 8.57%
7. Baylor – 8.82%
8. Nebraska – 8.89%
9. Texas – 9.23%
10. Kansas State – 10.45%
11. Iowa State – 10.71%
12. Kansas – 11.25%
So if we combine those two figures two the conference average, we can come up with some sort of rough “Value Over Average” (VOA) number. If you really want to know the formula I used for this, let me know...otherwise I’m thinking that a boring explanation of it would just distract from the numbers themselves.
Sack Rate VOA – O-Line
1. Texas Tech – 1.41
2. Oklahoma – 1.17
3. Missouri – 1.14
4. Oklahoma State – 1.11
5. Texas A&M – 0.98
6. Kansas – 0.93
7. Texas – 0.92
8. Baylor – 0.92
9. Colorado – 0.88
10. Kansas State – 0.86
11. Nebraska – 0.86
12. Iowa State – 0.82
Thoughts:• This really might be a pretty decent measure, as it doesn’t seem to give any benefit to the more elusive QB’s in the conference, like Bernard Jackson. Graham Harrell’s as stationary as they come—though he also gets rid of the ball quicker than anybody in the world—and Tech was far and away #1.
Okay, so now what happens if we combine the two “VOA” figures? Despite all of the limitations I mentioned above, this would still give us a decent read of who did the best job in combination run/pass blocking, right?
• I think it’s safe to say that Kansas was horrid at protecting from blitzes. On first and second downs, they were as good at avoiding the sack as anybody (outside Lubbock, anyway). On third and fourth downs, they were worse than Iowa State. Guess that might say something about the abilities of Kerry Meier, Adam Barmann, and Todd Reesing to make quick decisions, but it also says something about the O-line.
In this case, I guess you could say 2.00 = conference average, since it’s combining two averages.
Combined VOA – Offensive Line
1. Texas Tech – 2.32
2. Oklahoma State – 2.26
3. Oklahoma – 2.15
4. Kansas – 2.13
5. Missouri – 2.10
6. Texas A&M – 2.03
7. Texas – 2.02
8. Colorado – 1.97
9. Nebraska – 1.96
10. Kansas State – 1.81
11. Iowa State – 1.70
12. Baylor – 1.55
I think this does a pretty good job of comparing a team’s O-line talent to how it matches up with the offensive system. I’m sure Oklahoma (#3) and Texas (#7) had more NFL talent than Tech (#1) and OSU (#2) last year, but that didn’t allow them to do a better job. I do find ATM’s standing (#6) interesting considering how good people seemed to think they were this offseason. These numbers suggest that ATM’s rushing performance last year was due much more to the McGee-Goodson-Lane attack than the performance of the O-line...which makes sense to me. Mizzou didn’t have a problem with ATM’s running game last year until having to tackle Lane and chase Goodson 90 times (it felt like 90, anyway) caught up with them.
Next time, I’ll look at D-line performance using the same approach.
Friday, September 21, 2007
Stats Stats Stats: 2006 (O-Line)
Wednesday, September 5, 2007
Stats Stats Stats: 2006 (WR/TE)
When we last left off last week, we’d perused 2006 data for Big 12 RB’s in regard to success rates, et al. Now we’ll do the same thing for WR’s and TE’s.
Again, all of the following data is from only Big 12 games, and only from circumstances in which the score of the game is within 17 points. To qualify for the list, you have to have caught at least 10 passes.
I will look at the same criteria I did for RB’s:
1. Total number of successful plays. This simply measures consistency throughout the course of a season. For Big 12 WR’s that fit the criteria, the average was 19.1.There were 33 Big 12 WR’s that fit the above criteria, and here they are (those in bold are 2007 returnees):
2. Success Rate (% of plays in which they touched the ball—rushing or receiving—that resulted in a success, as defined in last week’s QB post). This measures effectivness and efficiency. The Big average in this category was 80.4%...as expected, much higher than that of RB’s (46.6%). Considering that almost half of all passes are incomplete (and therefore unsuccessful), this seems pretty logical.
3. Average % of Success (comparing what they gained on each play as a percentage of what was defined as successful for the given play...a 6-yard gain on 3rd-and-4 would be 150% success. I cap the success at 1000% for any given play). This measures playmaking and explosiveness. For WR’s, the average was 278%, compared to a 142% average for RB’s.
1. Adarius Bowman, Oklahoma State (33 successful plays / 91.7% success rate / 382% average success)Thoughts...
2. Joel Filani, Texas Tech (44 / 89.8% / 320%)
3. Maurice Purify, Nebraska (23 / 88.5% / 364%)
4. D’Juan Woods, Oklahoma State (25 / 92.6% / 310%)
5. Malcolm Kelly, Oklahoma (39 / 88.6% / 287%)
6. Will Franklin, Missouri (22 / 88.0% / 336%)
7. Quan Cosby, Texas (24 / 92.3% / 263%)
8. Todd Blythe, Iowa State (15 / 93.8% / 323%)
9. Jarrett Hicks, Texas Tech (18 / 85.7% / 323%)
10. Limas Sweed, Texas (19 / 73.1% / 321%)
11. Dominique Zeigler, Baylor (25 / 73.5% / 280%)
12. Robert Johnson, Texas Tech (33 / 80.5% / 253%)
13. Terrance Nunn, Nebraska (21 / 78.6% / 275%)
14. Alvin Barnett, Colorado (12 / 92.3% / 338%)
15. Yamon Figurs, Kansas State (16 / 76.2% / 318%)
16. Manuel Johnson, Oklahoma (24 / 88.9% / 200%)
17. Trent Shelton, Baylor (15 / 71.4% / 304%)
18. Tommy Saunders, Missouri (16 / 84.2% / 251%)
19. Danny Amendola, Texas Tech (17 / 70.8% / 293%)
20. Jared Perry, Missouri (20 / 76.9% / 233%)
21. Daniel Gonzalez, Kansas State (10 / 76.9% / 368%)
22. Jordy Nelson, Kansas State (13 / 76.5% / 292%)
23. Todd Peterson, Nebraska (12 / 85.7% / 271%)
24. Chad Schroeder, Texas A&M (18 / 72.0% / 228%)
25. Joachin Iglesias, Oklahoma (14 / 73.7% / 252%)
26. Thomas White, Baylor (13 / 76.5% / 251%)
27. Dexton Fields, Kansas (18 / 78.3% / 185%)
28. Billy Pittman, Texas (14 / 66.7% / 239%)
29. Eric Morris, Texas Tech (10 / 62.5% / 274%...was part-time RB)
30. Brad Ekwerekwu, Missouri (11 / 61.1% / 255%)
31. Brian Murph, Kansas (12 / 54.5% / 210%)
32. Earvin Taylor, Texas A&M (11 / 73.3% / 172%)
33. Jon Davis, Iowa State (13 / 86.7% / 61%)
• D’Juan Woods was a forgotten man last year at OSU after Adarius Bowman emerged the way he did, but it looks like he played a vital role. Bowman was a big-play guy, but Woods always managed to get what was necessary when teams keyed on AB. Against UGa, no WR emerged to fill that role—TE Brandon Pettigrew did his best, but if he’s the #2 option, then a chunk of OSU’s explosiveness is negated.
• Tech’s WR’s are such a nameless attack overall that I didn’t realize Filani was that much more accomplished than Hicks, Johnson, Amendola, etc.
• Todd Blythe caught only 15 passes in Big 12 play when the score was within 17 points, while the entirely ineffective Jon Davis caugh 13. That really does suggest that the best defense against Blythe is having Bret Meyer at QB. The dude’s 6’5, and you have no other weapons. Even if he’s double-covered, just throw a jump ball. It’s the only hope you have on offense.
• Jordy Nelson was criminally underused last season, though the case for his being injured (when he was thrown the ball last season, he was less effective than the mediocre Daniel Gonzalez) seems pretty strong here. And seeing how many times Josh Freeman threw him the ball last weekend, the case for his being healthy now is quite strong.
• If Will Franklin hadn’t been hurt for most of the last two games, he’d have almost certainly ended up #3 on this list.
• Since Limas Sweed has been the #1 guy at UT for something approaching 17 years now, it was surprising to see that a) Quan Cosby seemed to be the Go-To guy last year (possibly because Sweed was double-covered?), and b) when Sweed did catch the ball, he had a success rate lower than Yamon Figurs and Joachin Iglesias.
Now, on to TE’s. I bumped the critera for TE’s down to 7 touches (the other criteria stayed the same) so we could end up with at least 10 TE’s on the list. The average figures for TE’s on this list were...
Successful Plays: 13.7 Success Rate: 80.1% Average % of Success: 263%
Here are the TE’s who made the list:
1. Martin Rucker, Missouri (21 / 91.3% / 269%)Thoughts...
2. Brandon Pettigrew, Oklahoma State (11 / 84.6% / 312%)
3. Martellus Bennett, Texas A&M (21 / 84.0% / 264%)
4. Chase Coffman, Missouri (25 / 75.8% / 280%)
5. Jeron Mastrud, Kansas State (10 / 100.0% / 345%)
6. Joe Jon Finley, Oklahoma (10 / 83.3% / 260%)
7. Jermichael Finley, Texas (11 / 78.6% / 228%)
8. Derek Fine, Kansas (10 / 90.9% / 233%)
9. Rashaad Norwood, Kansas State (11 / 61.1% / 216%)
10. Riar Geer, Colorado (7 / 58.3% / 229%)
• Assuming Rashaad Norwood is unsuspended at some point, all 10 of these TE’s return in 2007. That’s impressive in and of itself.
• After railing against Martellus Bennett for most of the offseason, I’ll eat a smidge of crow here, ahem, and mention that his numbers were quite comparable to Rucker and Coffman. But they still weren’t better!
• Norwood’s suspension wouldn’t seem to hurt much since Mastrud caught almost as many passes and was far more effective. Mastrud caught four passes against Auburn, a solid start.
• As I mentioned earlier, Pettigrew was a big target for Bobby Reid in the opening game, and I’m thinking he’ll end up with more than 11 successful plays in 2007.
• The wildcard for 2007 is OU’s Jermaine Gresham, who is full speed (he was coming off of a knee injury last year) and running like a WR. I doubt any conference can even hold a candle to the TE production put out by the Big 12 this year.
Thursday, August 30, 2007
Stats Stats Stats: 2006 (RB's)
Yesterday, it was QB’s. Tonight, we take on running backs.
For RB’s, I will look at three criteria:
1. Success Rate (% of plays in which they touched the ball—rushing or receiving—that resulted in a success, as defined in yesterday’s QB post). This measures effectivness and efficiency.By looking at these three things together, one can get an idea of who the better/best RB’s in the conference were in Big 12 play. And again, all of the following data is from only Big 12 games, and only from circumstances in which the score of the game is within 17 points. To register on the list, you have to have touched the ball at least 10 times in these circumstances.
2. Average % of Success (comparing what they gained on each play as a percentage of what was defined as successful for the given play...a 6-yard gain on 3rd-and-4 would be 150% success. I cap the success at 1000% for any given play). This measures playmaking and explosiveness.
3. Total number of successful plays. This simply measures consistency throughout the course of a season.
There were 35 Big 12 RB’s who fit these criteria. Combining their ranking in each of the three criteria above, I came up with one all-encompassing ranking. Players in bold are returnees for 2007.
1. Jon Cornish, Kansas (75 successful plays, 47.5% success rate, avg% of success = 178%).
2. Michael Goodson, Texas A&M (53 / 54.6% / 171%)
3. Jorvorskie Lane, Texas A&M (63 / 56.8% / 143%)
4. Ryan Kock, Iowa State (31 / 59.6% / 169%)
5. Jamaal Charles, Texas (51 / 56.0% / 140%)
6. Shannon Woods, Texas Tech (52 / 48.2% / 159%)
7. Dantrell Savage, Oklahoma State (44 / 46.81% / 161%)
8. Leon Patton, Kansas State (44 / 46.8% / 161%)
9. Brandon Jackson, Nebraska (72 / 44.4% / 134%)
10. Cody Glenn, Nebraska (25 / 56.8% / 165%)
11. Keith Toston, Oklahoma State (29 / 59.2% / 148%)
12. Allen Patrick, Oklahoma (67 / 47.9% / 120%)
13. Chris Alexander, Texas A&M (15 / 83.3% / 256%)
14. Tony Temple, Missouri (39 / 44.8% / 155%)
15. Marlon Lucky, Nebraska (30 / 44.1% / 161%)
16. Paul Mosley, Baylor (31 / 44.3% / 143%)
17. Hugh Charles, Colorado (37 / 47.4% / 129%)
18. Selvin Young, Texas (40 / 44.4% / 125%)
19. Chris Brown, Oklahoma (27 / 46.6% / 145%)
20. Jason Scales, Iowa State (22 / 52.4% / 106%)
21. Jacob Gutierrez, Oklahoma (14 / 43.8% / 142%)
22. Mike Hamilton, Oklahoma State (15 / 38.5% / 134%)
23. Adrian Peterson, Oklahoma (17 / 39.5% / 125%)
24. Mell Holliday, Colorado (18 / 33.3% / 134%)
25. James Johnson, Kansas State (24 / 32.4% / 113%)
26. Earl Goldsmith, Missouri (9 / 42.9% / 119%)
27. Stevie Hicks, Iowa State (11 / 37.9% / 121%)
28. Byron Ellis, Colorado (9 / 42.9% / 106%)
29. Brandon Whitaker, Baylor (10 / 38.5% / 84%)
30. Brandon McAnderson, Kansas (6 / 28.6% / 145%)
31. Courtney Lewis, Texas A&M (9 / 32.1% / 88%)
32. Thomas Clayton, Kansas State (8 / 35.3% / 75%)
33. Josh Johnson, Iowa State (4 / 28.6% / 90%)
34. Kenny Wilson, Nebraska (5 / 33.3% / 53%)
35. Jimmy Jackson, Missouri (2 / 14.3% / 51%)
Here are the Top 5 in each category.
Successful Plays
1. Jon Cornish (75)
2. Brandon Jackson (72)
3. Allen Patrick (67)
4. Jorvorskie Lane (63)
5. Michael Goodson (53)
Success Rate
1. Chris Alexander (83.3%) – fullback
2. Ryan Kock (59.6%) – part-time fullback
3. Keith Toston (59.2%)
4. Cody Glenn (56.8%)
5. Jorvorskie Lane (56.8%)
Average % of success
1. Chris Alexander (256%)
2. Jon Cornish (178%)
3. Michael Goodson (171%)
4. Ryan Kock (169%)
5. Cody Glenn (165%)
So what does this tell us?
- Jimmy Jackson really didn’t have a very good year in 2006.
- Speed is good, but having a big back (Alexander, Kock, Glenn, Lane) can still pay off in the effectiveness category. Need 4 yards on 1st-and-10? Need 3 yards on 2nd-and-4? A big back’s your guy. Not that Mizzou would know anything about that. Sorry...I kid, I kid...
- Teams really geared up on Adrian Peterson. That, and he got hurt in the second Big 12 game, so his sample size is only from the Texas and Iowa State games.
- Jon Cornish really did make a case for being the best RB in the conference. Sometimes teams just run the same guy a million times, and that’s how he leads the conference in rushing. Cornish was consistent and effective, and he had some explosiveness as well.
- A&M’s running attack was (and will be in ’07) really effective. Between these numbers and the fact that Stephen McGee was the most effective 4th quarter QB in the conference, you can definitely see how their gameplan took shape.
- Just like the competition between Kerry Meier and Todd Reesing for KU QB has been kind of baffling, just as baffling is the ‘battle’ between Leon Patton and James Johnson for KSU RB. For a big back, Johnson seemed less effective in short yardage (just ask Tommy Chavis, who stoned him on 4th-and-1 in Columbia), and while he had some long runs, Patton had more.
- Suddenly Dave Matter putting Michael Goodson at the top of his RB ratings for 2007 makes a lot more sense.
Wednesday, August 29, 2007
Stats Stats Stats: 2006 (QB’s)
So I’ve finally finished entering play-by-play information from 2006...just in time for 2007 to begin. But that’s not going to stop me from looking into numbers to see what I find.
All of the following data is from only Big 12 games, and only from circumstances in which the score of the game is within less than 17 points. What’s the point of gauging what a player or team does when the game’s not in question? Instead of elaborating further, let’s just get started...
First, we’ll look at QB’s. As you’ll see a few posts from now, I will look at “Team Defense” in the same way as I do QB’s. I could do a “Team Offense” number as well, but a) that would be repetitive, and b) the QB is so important that a team’s offense is more greatly affected by his presence (or absense) than any other player’s. In other words, if I’m looking at individual QB’s, I’m not seeing the need to also look at Team Offense.
QB Success Rate
The good people at Football Outsiders (authors of the lovely Football Prospectus books) define a ‘successful play’ or ‘quality play’ by the following criteria:40% of the total needed on 1st down (i.e. if it’s 1st-and-10, 4 yards = success)
QB Success Rate will look at simply the % of plays that were ‘successful’ for each QB, no matter whether the play was a run or pass. This is an attempt to look at the overall efficiency with which a given QB runs his offense. I’m including runs because the effectiveness and efficiency of the running game is directly affected by the effectiveness and efficiency of the QB and his passing game.
70% of the total needed on 2nd down (if it’s 2nd-and-10, 7 yards = success)
100% of the total needed on 3rd or 4th down
Here are the QB Success Rates for 2006 Big 12 Play (for those with at least 10 plays):
1. Colt McCoy, Texas (49.12%, 399 plays)
2. Kerry Meier, Kansas (46.93%, 228 plays)
3. Zac Robinson, Oklahoma State (46.58%, 73 plays)
4. Bobby Reid, Oklahoma State (46.46%, 396 plays)
5. Graham Harrell, Texas Tech (46.21%, 448 plays)
6. Stephen McGee, Texas A&M (46.01%, 589 plays)
7. Chase Daniel, Mizzou (45.74%, 470 plays)
8. Paul Thompson, Oklahoma (45.20%, 542 plays)
9. Todd Reesing, Kansas (42.86%, 42 plays)
10. Jevan Snead, Texas (42.42%, 66 plays)
11. Bret Meyer, Iowa State (40.68%, 381 plays)
12. Zac Taylor, Nebraska (40.41%, 584 plays)
13. Shawn Bell, Baylor (39.93%, 268 plays)
14. Bernard Jackson, Colorado (38.18%, 385 plays)
15. Josh Freeman, Kansas State (37.93%, 406 plays)
16. Blake Szymanski, Baylor (34.58%, 107 plays)
17. Adam Barmann, Kansas (31.03%, 174 plays)
18. Dylan Meier, Kansas State (28.57%, 42 plays)
What a strange list. I mean...STRANGE. A few comments...
• Notice anything interesting about the top of the list? Kerry Meier? Really? Now...a majority of Kansas’ “successful” plays last year were due to Jon Cornish, but...Cornish was the RB when Todd Reesing and Adam Barmann were behind center too, and that didn’t stop their success rates from sucking. I realize Meier was injury-prone and had some very untimely turnovers last year, but...a) he was only a redshirt freshman, and b) he still moved the ball better than the other two. I just can’t see how naming Todd Reesing the starter is anything but counter-productive for the future of Kansas football. I’m not sure Kerry Meier’s anything but a slightly-glorified Kirk Farmer, but he’s still a better option than Reesing, I would think.
• Adam Barmann’s horrid numbers make his explosion against Nebraska that much stranger. Gonna have to go with the "Blind Squirrel and Nut" theory there.
• Yes, that’s Big 12 Offensive Player of the Year Zac Taylor finishing behind Todd Reesing, Jevan Snead, and Bret Meyer. Taylor was good for 1-2 bombs to Purify or others a game (and they obviously still won games), but as a consistent game manager, it appears that Taylor was a bit lacking.
• I think it’s pretty easy to see why Baylor’s QB job was up for grabs once again this offseason.
• I think it’s pretty easy to see why a redshirt freshman (Coach’s Son Hawkins) might be a better option for Colorado.
QB Success Rate by Quarter
Same as above, broken out by quarter. At least 10 plays needed once again. I’ll just do Top 8 by quarter
First Quarter
1. Paul Thompson (56.41%)
2. Jevan Snead (50.00%...on just 12 plays, all against K-State)
3. Chase Daniel (49.07%)
4. Colt McCoy (48.03%)
5. Graham Harell (46.45%)
6. Zac Taylor (45.22%)
7. Bernard Jackson (42.34%)
8. Blake Szymanski (40.91%)
Second Quarter
1. Graham Harrell (60.19%)
2. Kerry Meier (57.65%)
3. Stephen McGee (53.25%)
4. Colt McCoy (52.46%)
5. Bobby Reid (50.00%)
6. Bret Meyer (41.59%)
7. Chase Daniel (41.07%)
8. Zac Taylor (41.06%)
Third Quarter
1. Chase Daniel (52.53%)
2. Colt McCoy (52.22%)
3. Shawn Bell (52.00%)
4. Todd Reesing (50.00%...on just 16 plays)
5. Paul Thompson (48.12%)
6. Bret Meyer (46.46%)
7. Zac Robinson (45.83%)
8. Bobby Reid (43.33%)
Fourth Quarter (I’m listing all of them here, with total plays)
1. Bobby Reid (55.21%, 96 plays)
2. Stephen McGee (50.31%, 159 plays)
3. Zac Robinson (48.39%, 31 plays)
4. Jevan Snead (47.37%, 19 plays, all against K-State)
5. Kerry Meier (45.45%, 22 plays)
6. Todd Reesing (42.11%, 19 plays)
7. Josh Freeman (41.12%, 107 plays)
8. Adam Barmann (40.35%, 57 plays)
9. Colt McCoy (40.00%, 60 plays)
10. Paul Thompson (39.42%, 104 plays)
11. Chase Daniel (38.78%, 98 plays)
12. Zac Taylor (37.31%, 134 plays)
13. Shawn Bell (37.04%, 54 plays)
14. Bernard Jackson (36.67%, 60 plays)
15. Graham Harrell (35.29%, 85 plays)
16. Bret Meyer (34.43%, 61 plays)
Poor Blake Szymanski only had 5 snaps in the 4th quarter when games were within 17 points. This despite playing over half the Big 12 season. Ouch.
Oh, and just so you can compare...
Overall Success Rates by Quarter
First Quarter: 43.33%
Second Quarter: 43.24%
Third Quarter: 43.10%
Fourth Quarter: 41.84%
Now...I’m not totally sure what this says—and I realize we’re dealing with a lot of small sample sizes here—but I’m fascinated.
Comparing their success rate to what’s expected, here are the major QB’s and their rates by quarter...
• Chase Daniel: Q1 = 1.13 (his rate divided by Big 12 rate), Q2 = 0.95, Q3 = 1.22, Q4 = 0.93. Fantastic in the first and third quarters and below average in the second and fourth. Among many other things, this does highlight the strengths and weaknesses of Mizzou’s offensive coaching, something of which we’ve already had our suspicions. They were great in the first quarter, when plays were scripted, and in the third quarter when they were able to make adjustments at halftime. But once the scripts have expired, the play-calling gets a little shakier.
• Meanwhile, Stephen McGee (Q1 = 0.87, Q2 = 1.23, Q3 = 0.91, Q4 = 1.20) was the exact opposite. I thought that might say something about how more physical offensive attacks (like ATM’s) wear opponents down as the half progresses, but Graham Harrell’s strange line (Q1 = 1.07, Q2 = 1.39, Q3 = 0.93, Q4 = 0.84) might prove me wrong. Or does it prove me right? Harrell’s second halves were significantly worse than his first halves. I’m all for the spread offense, but it does show, I guess, that to win with the spread, you better put quite a few points up early because the defense probably won’t wear down against you.
• Paul Thompson’s line (Q1 = 1.30, Q2 = 0.81, Q3 = 1.12, Q4 = 0.94) shows that OU tended to assert itself early, then rely on its defense to pull through. Looking back at certain games (Missouri, Nebraska in the Big 12 title game), that sounds about right.
• For some reason, while McGee thrived in the fourth quarter, the QB’s for the other top teams in the conference (Taylor, McCoy, Thompson) were all below average in the fourth quarter. Part of that is sample size (the QB’s of crappy teams didn’t take nearly as many fourth quarter snaps in close games, so a few good plays could significantly affect their totals), but only part of it. Can’t really explain that.
Here’s the last thing I’m going to look at today...
QB Success Rate in Red Zone
Same as above, only limited by plays taking place inside the opponent’s 25 yard line. I use the 25 instead of the 20 because a) in OT, you get the ball at the 25, and since every play in OT is vita, I’m thinking your odds off scoring had better be just as good from the 22 as it is from the 18.
Conference-wide average was 43.84%.
1. Kerry Meier (60.98%, 41 plays)
2. Bobby Reid (58.06%, 62 plays)
3. Todd Reesing (54.55%, 11 plays)
4. Paul Thompson (53.85%, 91 plays)
5. Bret Meyer (47.22%, 72 plays)
6. Stephen McGee (45.38%, 130 plays)
7. Zac Taylor (44.44%, 90 plays)
8. Zac Robinson (42.86%, 21 plays)
9. Colt McCoy (42.39%, 92 plays)
10. Shawn Bell (41.94%, 62 plays)
11. Chase Daniel (41.24%, 97 plays)
12. Graham Harrell (40.58%, 69 plays)
13. Josh Freeman (32.76%, 58 plays)
14. Adam Barmann (31.03%, 29 plays)
15. Bernard Jackson (23.73%, 59 plays)
A couple interesting things about this...
• First, it does show that the spread offense isn’t amazingly effective in the red zone—Daniel, Harrell, and Bell/Szymanski were all relatively unimpressive.
• However, it doesn’t show that a particular style of offense works better than another. McGee and his ‘cloud of dust’ offense, fared only marginally better than the spread QB’s. KU was great in the red zone (at least with Meier and Reesing at the helm...not so much Barmann), which makes sense because of Jon Cornish. However, why was Cornish effective when other big RB’s (Lane/Goodson at ATM, Brandon Jackson at NU) were just slightly above average? Or in the case of ISU's hoss Ryan Kock, quite below average? (That one can be explained by the fact that Kock only played well against Mizzou...ahem).
Alright, that’s enough for today. I’m going to go position-by-position over the next week or two, pointing out interesting things I find. It stinks that I couldn’t have finished entering the data earlier in the offseason, but...oh well. Better late than never.
Tomorrow: RB’s.
Wednesday, August 15, 2007
Opinions needed
So I'm finally almost through with entering play-by-play info from the 2006 football season, and some time toward the end of next week I'll be compiling some stats to look at before the new season begins. I've been looking at some basic Football Outsiders stats like Line Yards and quality plays, but I'm wanting to figure out a way to weight the costliness of turnovers. I'm not a huge fan of "points off turnovers" for football because it doesn't tell the whole story. For one thing, how close your team was to scoring before the turnover seems just as important as whether your opponent scored. For another, there's no telling how much your turnover actually led to an opponent's points. If you fumbled at your 1, and they scored a TD on the next play, then yes...your turnover played a key role. But if you threw an INT at your opponents' 20, and they just happened to drive 80 yards for a TD, that doesn't say a lot. And if you threw an INT at your opponents' 20, and they drove 30 yards before punting, that's still a pretty costly turnover because you blew a key scoring opportunity.
As I've been entering the play-by-plays, I came up with a pretty primitive equation that simply factors where the turnover took place (3 points if it took place between either your or your opponent's goalline and 20, 2 if it took place between either 20 and 40, 1 if it took place between the 40's) and how close the game was at the time of the turnover (2 points if within two possessions, 1 point if within 24 points, 0 if over 24 points). There's a maximum of 5 points available. If the turnover is returned for a TD, it's an automatic 5 points. That's okay, but it still doesn't really tell the story. For instance, if you fumbled at your 45 and it was returned to your 1, that's 1 point because it took place at your 45. I guess I could simply look at resulting field position to fix that problem...but that would open up the opposite problem (you fumble at your opponent's 1, they return it to the 45...and it's only 1 point). Maybe an average of the two?
Anyway...for all you lurkers out there (our readership has gone up nicely, but the volume of comments has not), any ideas on this? Is there a better way to measure turnover costliness? I'm open to trying anything that can fit in an excel equation.
(Also, if you lurkers out there just wanna say hi, that's cool too.)
Friday, May 25, 2007
Numbers Numbers Numbers: ‘Defensive Playmakers’ Edition
Well, I have officially finished entering play-by-play information for Mizzou’s 2006 football games. I endeavor to get all Big XII teams entered by the end of the summer, but in the meantime I thought I’d start looking into what the data is telling me so far.
As I mentioned to Dave Matter earlier this week, one piece of information I’ve investigated is “successful” plays.
[T]he writers of Football Prospectus (http://www.footballoutsiders.com/) define a "successful play" as...As I entered play-by-play, I looked a few extra variables. Was the play a ‘success’ as defined above? What % of the first down did the play gain? What % of ‘success’ did the play gain? Et cetera. With this information, I thought it would be interesting to look at who the best “playmakers” were on the 2006 Mizzou squad.
- Gaining 40% of needed first down yardage on 1st down
- Gaining 70% of needed first down yardage on 2nd down
- Gaining 100% of needed first down yardage on 3rd or 4th down
This post will only look at defense.
For defense, a “successful play” would obviously be a play that prevented the offense from reaching the ‘success’ defined above. I also divided the plays into two categories: 1) plays when the score of the game was ‘close’ (i.e., the margin was somewhere between 0-15 points), and 2) plays when the score of the game was not close. I figure the best playmakers are the ones who earn their stats in closer games.
So what did I find? Well, as would be expected, using the standards above, defensive linemen are going to have a higher percentage of ‘successful’ plays simply because they start closer to the line of scrimmage than LB’s or DB’s. On the flipside, safeties are, by name, guys who are more likely to stop a big play than make a big play—and therefore, they’ll have the lowest percentage of ‘successful’ plays. The average Mizzou D-lineman had a ‘Successful Play to Non-Successful Play’ Ratio of 2.18. Defensive backs? 0.37. So if I’m trying to figure out a measure to use for any defensive player, I have to take their position into account. A defensive back with a 1.00 ratio is more effective than a defensive lineman with a 2.00 ratio. Makes sense, right?
So, for starters, I looked at what I’m tentatively calling the Position Playmakers Ratio. I’d call it the PPR, but giving it an acronym would make it seem like I’m sold on the name more than I actually am. I’m open to suggestion on the name.
What this ratio takes into account are two things: 1) the ‘Successful Play to Non-Successful Play’ Ratio (in ‘Close’ situations only...meaning, plays made when the score is within 15 points) mentioned above for a given player, and 2) the average Ratio for a player at that position. I basically just divide (1) by (2). If the ratio is above 1.0, then the player’s (1) is higher than the average player at their position. For now, I’ve only inputted Mizzou’s numbers, so (2) is only the average for Mizzou players. As I get more teams entered, (2) will change, and it will become an effective comparison of players from different teams.
So anyway, what are the PPR’s...I mean, Position Playmakers Ratios for last year’s Mizzou defensive players?
Lorenzo Williams – 1.95Those are the 8 players who were above 1.00, meaning they had a better (1) than the average at their position.
Dedrick Harrington – 1.56
Xzavie Jackson – 1.35
Darnell Terrell – 1.32
Pig Brown – 1.25
Brian Smith – 1.19
Stryker Sulak – 1.17
William Moore – 1.03
David Overstreet – 0.96Now, there are plenty of flaws in this measure. It doesn’t take into account Takeaways, which are obviously the most effective way of ‘playmaking’. If we figured out a way to take INT’s, forced fumbles and fumble recoveries into account (and I’ll get there eventually), Marcus Bacon (2 INT’s, 5 FF’s, 3 FR’s) would move way up the list, as would Lorenzo Williams (2 FF’s, 3 FR’s, plus 2 blocked kicks!) and Stryker Sulak (1 FF, 3 FR’s).
Marcus Bacon – 0.91
DeMarcus Scott – 0.79
Brandon Massey – 0.76
Jaron Baston – 0.75
Jamar Smith – 0.67
Ziggy Hood – 0.63 *
Hardy Ricks – 0.57
Brock Christopher – 0.53
Tommy Chavis – 0.51
Domonique Johnson – 0.45
* It’s worth noting that, in the first two games of the season, before he got hurt, 100% of Ziggy Hood’s tackles in ‘close’ games resulted in a success for the defense. Of course, there weren’t too many plays to make in ‘close’ situations against Ole Miss or Murray State since they became blowouts relatively quickly, but I think this does show that Hood was very obviously not the same player after his broken foot.
There is also another aspect to take into account regarding ‘playmaking’: in what percentage of your team’s plays are you involved? Success or no, close game or no, in what percentage of your team’s plays do you get a tackle or solo tackle? For this number, you’d expect LB’s to have a much higher % than DT’s or CB’s (you’d HOPE your CB’s aren’t too high), so let’s once again create a ratio out of the tackles a player made compared to what would be expected out of his position.
Taking injuries (and therefore fewer tackle opportunities) into account, here are the Top Ten Mizzou players in what we’ll call (for now) Tackles/Position Ratio (TPR?):
Darnell Terrell – 1.91You can tell pretty quickly that the players on this list were the ones who were likely on the field more—all these players started a good chunk of the season. So for now, this seems like as much a measure of tackle opportunities as it does anything else. However, considering how heavy the rotation was among the safeties—Overstreet, Massey, W. Moore, and Pig Brown—it is relatively impressive how high Overstreet’s ratio is here.
Xzavie Jackson – 1.59
David Overstreet – 1.55
Ziggy Hood – 1.47
Lorenzo Williams – 1.38
Brian Smith – 1.21
Marcus Bacon – 1.15
Stryker Sulak – 1.09
Brandon Massey – 1.06
Hardy Ricks – 1.06
And the fact that Terrell had that many more tackles than could be expected of the CB position is a bit confusing—did he have more tackles because the guy he was covering had more catches than they should have (he was, after all, almost always single-covering the opposition’s best WR), or because he was better in run support, or both?
For now, let’s say that having a high ‘Tackles/Position Ratio’ is a good thing (which is obviously debatable). What if we created one ‘playmaker measure’ by multiplying this TPR with the ‘PPR’ from above (what would we call it, the TPRPPR?)? The higher the number, the higher tackles and ‘successful plays’ than others at their position. Sounds like a great idea, right? Let’s see what that number tells us. Here are the top ten Mizzou players (bold = returning in 2007):
1. Lorenzo Williams – 2.69Lots of ground still to cover with these measures (and other ones I haven’t touched on yet), but I thought this would be something interesting to look at. The fact that Lorenzo Williams is that much higher than anybody else, I think, says something. What it says, exactly, will be more clear when I get other teams’ data entered. I bet you can’t wait.
2. Darnell Terrell – 2.51
3. Xzavie Jackson – 2.14
4. Dedrick Harrington – 1.63
5. David Overstreet – 1.50
6. Brian Smith – 1.43
7. Stryker Sulak – 1.28
8. Marcus Bacon – 1.04
9. Ziggy Hood – 0.93
10. William Moore – 0.90
As always, I open the floor here for feedback. Are there other things I should be looking at with this data?
Saturday, April 28, 2007
Numbers Numbers Numbers (Part Three)
Part One
Part Two
I’m no longer at a loss as I was in Part Two! Go me! Well...I still don’t totally know where this is going, but I’ve found some interesting things to share.
First, here’s what I’ve done since the last post:
- I’ve entered box scores for 2003. I now have four seasons’ worth of Big XII box scores.
- I’ve dabbled in some pivot tables (oh baby!).
- I’ve looked further into the predictive values of my numbers (i.e. which factors are precursors of possible success the next season).
What do you figure is the most important of this list: Rush Offense, Pass Offense, Rush Defense, or Pass Defense? In other words, if you had to build a team from scratch, where should you start? As always, it depends on how you define those terms, I guess. Are we talking yards or yards per play? We’ll look at both. For results between 2003 and 2006 (Big XII teams), here are the correlations between quality in one area and overall success.
First, let’s look at the importance of overall yardage:
1. Rushing Yards (correlation: 0.51)This makes sense, really. Pure yardage stats are extremely circumstantial—how many times you run or pass (and therefore how many yards you rack up) can be closely dictated by the game situation (you’ll run more, and therefore rack up more yards, if you’re ahead; you’ll pass more, and therefore rack up more passing yards, if you’re behind). So if you’re able to get up early, you’ll likely run the ball more. However, some teams pass more than others when the game is close, so high passing yards could mean a couple different things (especially when Texas Tech is a member of your conference).
2. Opponents’ Rushing yards (0.48)
3. Opponents’ Passing yards (0.24)
4. Passing Yards (0.14)
Just looking at yards isn’t a wonderful gauge because so many other game aspects/circumstances affect this category. Does it make more sense to look at yards per attempt?
1. Opponents’ Yards Per Pass Attempt (0.49)There is little disparity here. They’re all of roughly the same importance. This almost becomes as much a measurement of big plays as passing or rushing proficiency. I guess we could average the two.
2. Yards Per Rush Attempt (0.47)
3. Opponents’ Yards Per Rush Attempt (0.46)
4. Yards Per Pass Attempt (0.44)
1. Rush Offense (0.49)If this is a legitimate way of looking at things, then it would suggest that having a good passing game is the easiest thing to do, and it affects games the least. Meanwhile, having a strong running game is the hardest, and it has the most impact. Let’s look at the results to see if that’s true.
2. Rush Defense (0.49)
3. Pass Defense (0.37)
4. Pass Offense (0.29)
(To determine the “Best” for each of these, I just looked at yards per game and yards per attempt and used my own discretion)
2003Figuring that conference record is the best gauge (since there’s a huge disparity among non-conference schedules), the order might actually be 1) Run Defense, 2) Pass Defense, 3) Run Offense, 4) Pass Offense. But really, there’s just not a distinguishable difference here, at least among the first three categories. All we know for sure is a) Texas Tech is pretty good at throwing the ball (duh), b) the “defense wins championships” cliché has some statistical backing (even bigger duh), and c) Missouri screws these comparisons up by showing up three times and never having a great conference record (for Mizzou fans, the biggest duh of all).
Best Running Game: Missouri (conference record: 4-4, overall: 8-5)
Best Run Defense: Kansas State (7-2, 11-4, Big XII Champs)
Best Pass Defense: Oklahoma (8-1, 12-2)
Best Passing Game: Texas Tech (5-3, 8-5)
2004
Best Running Game: Texas (7-1, 11-1)
Best Run Defense: Oklahoma (9-0, 12-1)
Best Pass Defense: Missouri (seriously, look up the numbers...best in both yards and yards per attempt) (3-5, 5-6)
Best Passing Game: Texas Tech (5-3, 8-4)
2005
Best Running Game: Texas (9-0, 13-0)
Best Run Defense: Kansas (3-5, 7-5)
Best Pass Defense: Texas (9-0, 13-0)
Best Passing Game: Texas Tech (6-2, 9-3)
2006
Best Running Game: Oklahoma State (3-5, 7-6)
Best Run Defense: Texas (6-2, 10-3)
Best Pass Defense: Missouri (again, look up the numbers) (4-4, 8-5)
Best Passing Game: Texas Tech (4-4, 8-5)
Combined Records
Best Running Game - conference: 23-10 (.697 win %), 39-12 (.765 win %)
Best Run Defense – 25-9 (.735), 40-13 (.755)
Best Pass Defense – 24-10 (.706), 38-13 (.745)
Best Passing Game – 20-12 (.625), 33-17 (.660)
Yeah, that took a lot of words and numbers to basically reaffirm that you can win by being good at any number of things, but you’re more likely to win if you’re good at a number of things. I’m brilliant. So what do you think? I can twist these numbers to say that any aspect is more important than any other...what is your view of the most important? My assumption when I started was Running Game > Passing Game and Defense > Offense. Yours? Is there some other way I should look at this? I actually enjoy doing this for some nerdish reason...
Sunday, April 15, 2007
Numbers Numbers Numbers (Part 2)
Okay, let’s just get this out of the way right off: I’m at a loss right now. Anyway...
In case you didn’t read Part 1 of what will hopefully become a pretty lengthy series of Football Numbers posts (assuming I figure out where to go with this), here’s a quick summary of Part 1: last week I got things started by entering the 2006 box scores for Big XII teams and running some simple correlations to see what statistical categories had the biggest impact on wins and losses for each team. Each team had a completely different list, and I was curious if these correlations tended to change drastically from year to year or if each team had something of a blueprint.
(Also, I changed from running the basic Excel correlation, which is a Pearson’s correlation, to running a Spearman correlation, which is “bivariate normal.” Like the term ‘1080i’ when it comes to new big screens, you don’t have to know what “bivariate normal” is, you should just know that it’s cool.)
So what’s happened since then? Well, this week I entered the box scores from 2005 and 2004 with three themes in mind: 1) Would each team have similar key variables every year...in other words, is there something of a blueprint for each team? 2) Was a high/low correlation in any one variable a sign of a good/bad team (i.e. if I listed out all the correlations and tied them to win %, would there be a high correlation in any one area?)? 3) Was a high/low correlation in any one variable a predictor of success/failure the next season?
In the methods I used, I really didn’t find much. So I’ll be asking for some help. But first, I’ll summarize what I found...in all its non-glory...
Does each team have a blueprint?
In a word...no.
Well, that’s not quite right. In all, if you look at the average correlations for all the categories I entered, you’ll see that each team still has something of a blueprint—each team has different correlations and categories which most directly lead to their success/failure. However, correlations for any most variables vary significantly from season to season. For instance, Baylor’s #1 most important overall category is Opponents’ Yards per Carry (with an average correlation of -0.73). The correlation for this variable was -0.81 in 2004, -0.88 in 2005, and -0.49 in 2006. For Missouri, Team Yards Per Carry was the #3 most important variable overall (an average of 0.66). Its value was 0.62 in 2004, 0.92 in 2005, and 0.43 in 2006. Even the most important overall variables vary wildly from season to season.
This makes sense when you think about it, though. There’s always a change in personnel, and football, like other games, is a game of inches. So even if a team is perfectly consistent in its play-calling and relatively consistent in its execution, there are still all sorts of variables that factor into their success. This isn’t surprising.
It was disappointing, though. I’d love to stumble across a magic bullet, after all.
Anyway...with that in mind, here are the strongest correlations for each team (again, I ranked them according to absolute value—some high correlations were positive, some were negative).
BaylorPredictably, here are the Top 5 Variables:
1. Opponents’ Yards Per Carry (-0.73)
2. Opponents’ Rushing Yards (-0.66)
3. First Down Ratio (0.51)
4. Opponents’ Turnovers (0.51)
5. 3rd Down Ratio (0.47)
Colorado
1. Team Rushing Attempts (0.69)
2. Opponents’ Rushing Attempts (-0.65)
3. First Down Ratio (0.65)
4. Team Rushing Yards (0.64)
5. Opponents’ Rushing Yards (-0.58)
Iowa State
1. 3rd Down Ratio (0.65)
2. Opponents’ Yards Per Passing Attempt (-0.59)
3. Team Rushing Yards (0.54)
4. Team Rushing Attempts (0.54)
5. Yards Per Carry (0.53)
Kansas
1. 3rd Down Ratio (0.62)
2. Rushing Yards (0.52)
3. 3rd Down Conversion % (0.52)
4. Opponents’ Rushing Yards (-0.49)
5. Pass Completion % (0.49)
Kansas State
1. 3rd Down Ratio (0.73)
2. First Down Ratio (0.68)
3. Rushing Attempts (0.62)
4. Time of Possession (0.59)
5. Opponents’ Completion % (-0.58)
Missouri
1. Rushing Yards (0.72)
2. Opponents’ Rushing Yards (-0.68)
3. Yards Per Carry (0.66)
4. First Down Ratio (0.56)
5. First Downs (0.53)
Nebraska
1. Rushing Yards (0.60)
2. Turnover Margin (-0.60)
3. Opponents’ Yards Per Carry (-0.55)
4. Opponents’ Yards Per Passing Attempt (-0.55)
5. Team Rushes (0.55)
Oklahoma
1. First Down Ratio (0.64)
2. Opponents’ First Downs (-0.60)
3. Opponents’ Yards Per Passing Attempts (-0.56)
4. Opponents’ Yards Per Carry (-0.54)
5. Opponents’ Turnovers (0.51)
Oklahoma State
1. Rushing Yards (0.64)
2. Opponents’ Yards Per Passing Attempt (-0.60)
3. Passing Attempts (-0.59)
4. 3rd Down Ratio (0.59)
5. Opponents’ Rushing Yards (-0.55)
Texas
1. First Down Ratio (0.73)
2. Opponents’ First Downs (-0.72)
3. 3rd Down Ratio (0.70)
4. Opponents’ Total Plays (-0.63)
5. Opponents’ 3rd Down % (-0.61)
Texas A&M
1. Opponents’ First Downs (-0.65)
2. First Down Ratio (0.64)
3. Opponents’ Yards Per Carry (-0.59)
4. Opponents’ Rushing Yards (-0.58)
5. Opponents’ Turnovers (0.57)
Texas Tech
1. 3rd Down Ratio (0.71)
2. Yards Per Carry (0.66)
3. Opponents’ Yards Per Passing Attempts (-0.59)
4. Opponents’ 3rd Down % (-0.58)
5. Rushing Yards (0.55)
1. 3rd Down Ratio (0.62)So after all of this, we can come to one specific conclusion...yards matter.
2. First Down Ratio (0.61)
3. Team Rushing Yards (0.53)
4. Opponents’ Yards Per Passing Attempt (-0.53)
5. Opponents’ Rushing Yards (-0.51)
I know, what an amazing thought. I could have come up with that without looking at a single box score. Just as on-base percentage is the single most key variable in baseball (if you’re not making outs, you’re more likely to score points), yards are the single most key variable in football (if you’re advancing the ball, you’re more likely to score points). Brilliant. Moving on...
Was a high/low correlation in any one variable a sign of a good/bad team?
Okay, so I correlated statistics from individual games to the results of those games and found that the team that controls the ball, converts on 3rd downs, and ends up with more yards, probably ends up with more points. What happens if we take a step back and look at a season’s success instead of individual games? What happens if I take those individual game correlations and tie them to a team’s winning percentage for any given year?
In other words, instead of looking at what’s most important for winning a game, what’s most important for a winning season?
(I’m having trouble wording this right. I’m basically looking at the correlation between correlations and win %, but “correlations between correlations” doesn’t really sound all that clear, does it?)
Anyway, I didn’t find much. First of all, none of the resulting correlations were all that strong (the highest was 0.41), but here was the main conclusion I could draw: the more important Opponents’ Yards Per Pass Completion, Opponents’ Yards Per Pass Attempt, and overall Opponents’ Passing Yards are to you (i.e. if it has a higher-than-normal correlation), the worse your record is.
What does that mean? I’m honestly not sure. Does it have to do with big plays? In other words, if you have a high correlation in these categories, I guess that means you win or lose games depending on how many big passing plays you give up. That suggests you give up quite a few big plays, doesn’t it? That’s all I could come up with.
Of course, I don’t know how much thought this is worth, since a 0.41 correlation with a relatively small sample size isn’t all that telling.
Was a high/low correlation in any one variable a predictor of success/failure the next season?
I looked at this one the same way I looked at the last one, only instead of tying individual game correlations to a team’s win %, I tied them to the next season’s win %. I’m not working with a huge sample size here (2005’s win % with 2004 correlations and 2006’s win % with 2005 correlations...2 years for 12 teams), but here’s what I’ve come up with so far.
(And it should be noted the best predictor of next year’s win % is...this year’s win %. That doesn’t fill me with confidence in these weak correlations. Still no magic bullet.)
But for the sake of sharing, I did find something interesting. You know the ‘big play’ rule from above? “The more important Opponents’ Yards Per Pass Completion, Opponents’ Yards Per Pass Attempt, and overall Opponents’ Passing Yards are to you (i.e. if it has a higher-than-normal correlation), the worse your record is”? Well, looking at the next year’s numbers, I can say that if your own yards per completion are more important (i.e. have a higher correlation) in any given season, the worse your record is going to be the next season. This makes a little bit of sense, too.
And really, going back to the baseball stats analogy earlier, this is a lot like a team’s batting average with runners in scoring position (RISP). The ’03 Royals had a strangely high RISP (and won 83 games), and it gave fans (and the front office) an artificially inflated view of where the organization was as a whole. Well, in ’04 they were quite below average in that category, and they ended up losing 104 games. RISP balanced out in the end. They weren't especially good at getting hits with runners in scoring position; they just got on an extended hot streak. Which was followed by an extended cold streak.
Well, it appears that big pass plays are somewhat the same. If you’re giving up a lot of bombs one year, your record will probably suffer, but it will likely even out the next season, and those low-percentage passes won’t find the hands of opposing WR’s (and therefore significantly raise your opponents’ yards per catch and affecting the outcome of a game) quite as often.
Now, again...these correlations aren’t very high—there are other factors involved, like some teams being better than others at the fundamentals of tackling and/or pass coverage, and some teams just having more talent—but I don’t find it a total coincidence that the same categories emerged, inversely, when looking at a team’s record from year to year.
And just in case you’re wondering, here were the teams with the highest 2006 correlations in the category of Team Yards per Completion: 1) Texas, 0.70, 2) Texas Tech, 0.69, 3) Oklahoma State, 0.30, 4) Kansas State, 0.29. For the record, Missouri was at just 0.03, which hopefully suggests that those bombs that fell just out of the reach of WR’s last year (until the bowl game, anyway) will find a little more success. As for the team at the top of the list, I guess don’t be surprised if Colt McCoy’s reputation for throwing a great deep ball loses as smidge of its luster this year, even though most of his major receiving targets return for 2007.
Summary
So...what have I learned in the process of entering all this data? Not nearly as much as I would have hoped. But I do have all of this data, and I plan on compiling more. My question to you is, what should I do with it? And what data would be the best to compile? After I get back to about 2001, I figure I’ll move on to play-by-play data, but I don’t want to dig in that deep without some idea of what I’m looking for. So I’m asking for help from any burdgeoning data or sabermetrics nerd reading this...let’s make this a community project. Let me know where you think I should go with this.
Feel free to share any thoughts in the comments section.
Saturday, April 7, 2007
Numbers Numbers Numbers
I almost waited to post this because I was having too much fun with the K-State basketball situation, but...I guess I’ll move on. For now.
I’ve officially embraced my stat-nerddom like never before. In preparation for the upcoming 2007 football season, I’m diving into box scores and swimming around a bit. And I plan on doing this all summer. Let’s just say that, as much as I enjoy basektball and basketball stats, football ranks much higher on my list. You’ve been warned.
Here was my first item of business: look at 2006 box scores for all Big XII teams and do a simple correlation. What would happen if I looked at a bunch of different statistical categories? Which categories would have the highest correlation to actual success and failure for each team? Would the key categories be the same for every team? Absolutely, positively not.
WARNING: For those of you (ahem, The Beef) who begins to get a headache when statistics terms are discussed, please skip over the italicized portion below.
What is a correlation? From wikipedia, which has all the answers: "In probability theory and statistics, correlation, also called correlation coefficient, indicates the strength and direction of a linear relationship between two random variables. In general statistical usage, correlation or co-relation refers to the departure of two variables from independence, although correlation does not imply causation. In this broad sense there are several coefficients, measuring the degree of correlation, adapted to the nature of data." Crystal clear? Basically it means, if one variable goes up, another variable will likely go up too. Correlation isn't equal to causation, but the higher the correlation, the stronger the relationship between the two numbers.
A correlation coefficient can only be between -1 and 1. In all of this correlation analysis, keep in mind that I’m looking at the absolute values of these correlations. Just about any category involving opponents’ success probably had a negative correlation (i.e. the fewer yards for your opponents, the higher correlation to your team’s success), but being that I was looking for the strongest overall correlation, I looked at absolute values.
Okay, now that that's out of the way...
First, here were the statistical categories with the five highest correlations to success, conference-wide:
1. First Down Ratio (your total FD’s versus your opponent’s FD’s) (correlation: 0.68)Now, your first impression as I list those five are, DUH. Of COURSE first downs and third down conversions matter. Well, everything matters to some degree. There’s no denying that. If the top category had been rushing yards or opponents’ total yards or time of possession, you’d have said ‘duh’ to that too. That was kind of my idea in looking at this. Everything matters, but what matters the most to each team?
2. Opponents’ Total First Downs (0.59)
3. Third Down Conversion Ratio (your 3rd down converion rate versus your opponent’s) (0.53)
4. Opponents’ Yards Per Passing Attempt (0.51)
5. Opponents’ Third Down % (0.48)
Well, here’s a look at each team’s top five and what it probably suggests about each team.
There were a few weird things about Missouri’s numbers. First of all, as you’ll see, these correlation numbers are much higher than other teams’. In other words, these categories were more tied to Missouri’s success/failure than other teams’ highest categories were to theirs. Also, the 0.84 correlation for Missouri’s rushing attempts was just about the highest correlation on the board. Obviously there’s some cause-effect working there—Missouri is more likely to run the ball when they’ve already got the lead. But does that also mean they should run the ball more at any time in the game?Missouri
1. Rushing Attempts (0.84)
2. Time of Possession (0.72)
3. First Down Ratio (0.71)
4. Rushing Yards (0.70)
5. Opponents' Rushing Attempts (0.70)
One other weird thing: for every other team in the conference, time of possession meant next to nothing (which was surprising in and of itself). But for Mizzou, it was just about the most important thing. Gary Pinkel spends a lot of time talking about how TOP doesn’t matter—it’s total plays that matters. Well, total plays had a relatively high correlation (in the 0.6 range), but TOP meant more. I wasn’t expecting that.
BaylorTakeaways meant more to Baylor than any other team, and by a pretty wide margin. They needed to create some extra opportunities for themselves, and when they didn’t do it, they lost.
1. Opponents’ turnovers (0.74)
2. Opponents’ completion % (0.70)
3. Opponents’ 3rd down conversion attempts (0.69)
4. Opponents’ 3rd down conversion % (0.64)
5. Third Down Conversion Ratio (0.60)
ColoradoColorado’s correlations were the only team’s stronger than Mizzou’s, and all of these categories have to do with ball control...which makes sense. Colorado didn’t have many explosive weapons on offense, and their defense was good at preventing the big plays. Whoever was able to dictate the tempo won the game.
1. Opponents’ Rushing Attempts (0.87)
2. First Down Ratio (0.80)
3. Third Down Conversion Ratio (0.77)
4. Total Plays Ratio (0.71)
5. Opponents’ Total Plays (0.71)
Iowa StateTo everyone else in the conference, opponents’ passing attempts and completions were just about the least important categories on the list. However, for Iowa State it was extremely important. To me, that suggests that Iowa State had trouble building a lead, but when they had one (and their opponents therefore had to pass a lot), they were decent at holding onto it.
1. Third Down Ratio (0.79)
2. Opponents’ Passing Attempts (0.67)
3. Rushing Yards (0.66)
4. 3rd Down Conversion % (0.66)
5. Opponents’ Pass Completions (0.64)
KansasLike Colorado, ball control meant absolutely everything to Kansas. If their opponents were running the ball well and getting first downs, Kansas was screwed. However, if KU was able to string together some first downs, they were in good shape.
1. Opponents’ First Downs (0.67)
2. 3rd Down Conversion Rate (0.67)
3. Opponents’ Yards per Carry (0.60)
4. Opponents’ Total Plays (0.57)
5. First Down Ratio (0.54)
Kansas StateSense a trend here? Me too.
1. 3rd Down Conversion Ratio (0.86)
2. 3rd Down Conversion Rate (0.75)
3. 3rd Down Conversions (0.66)
4. Team Passing Attempts (0.57...this was a negative correlation)
5. Turnovers (0.55)
NebraskaThis tells me that Zac Taylor winning Big XII Offensive Player of the Year was an even bigger joke than I thought it was. If this team was running the ball well, Nebraska was winning. If they weren’t they were losing. And if they were running the ball well, that opened up the passing game. Zac Taylor was about the 9th most important player on the offensive side of the ball.
1. Rushing Yards (0.82)
2. 3rd Down Conversion Ratio (0.75)
3. Rushing Attempts (0.75)
4. Pass Completion % (0.72)
5. 3rd Down Conversion Rate (0.71)
Okay, maybe that was a bit overboard. But my point is valid, and you know it.
OklahomaThis was a pretty unique set of categories. We knew Paul Thompson wasn’t all that important to OU’s overall success—his one job was “Don’t screw up” and he did a decent job of that—but this pretty much verifies that the entire offense was assigned the same role. That’s pretty surprising considering that OU was a pretty strong rushing team...especially when that Peterson guy was healthy. However, when you think about it, OU’s improvement coincided with their defense’s significant improvement. It was a huge disappointment the first month or so of the season, but after the Texas game, things clicked, and OU didn’t lose again the rest of the season (to a team not named Boise State, anyway).
1. Opponents’ Completion % (0.74)
2. First Down Ratio (0.65)
3. Penalty Yards (0.61)
4. 3rd Down Conversion Ratio (0.58)
5. Opponents’ Rushing Yards (0.54)
Oklahoma StateThis makes sense to me. OSU’s offense was consistently good all year. Lots of explosiveness and big plays. However, the defense...not so good. When the defense—particularly the pass defense—stepped up, success followed.
1. Opponents’ Completion % (0.72)
2. First Down Ratio (0.65)
3. Opponents’ Yards Per Passing Attempt (0.63)
4. Opponents’ Rushing Yards (0.61)
5. Rushing Attempts (0.61)
TexasBig plays = good. Giving up big plays = bad.
1. Yards Per Pass Attempt (0.75)
2. Yards Per Pass Completion (0.73)
3. First Down Ratio (0.70)
4. Opponents’ Yards Per Passing Attempt (0.64)
5. Pass Completion % (0.63)
Texas A&MI have absolutely no idea what to make of this. Seriously. Do you? Um, ball control’s important, I guess?
1. Opponents’ Total First Downs (0.67)
2. First Down Ratio (0.65)
3. Opponents’ Yards Per Rush (0.62)
4. Opponents’ 3rd Down Attempts (0.57)
5. Opponents’ Pass Completion % (0.57)
Texas TechThis one makes sense too. Just like OSU (even moreso), the offensive yards were always there. They’re always going to get first downs, but if you get more than they do, you’re probably going to win. Also, when they’re ahead, they run more...just like Missouri.
1. First Down Ratio (0.81)
2. Third Down Ratio (0.76)
3. Rushing Yards (0.68)
4. Turnover Ratio (0.67)
5. Yards Per Pass Attempt (0.66)
Over the next couple of weeks—in the lead-up to the Black & Gold Game—I’ll be taking a look at each Big XII team, and I figure...you know...since I spent all this time looking at numbers, I should probably use them in those previews too, huh?
Again, you’ve been warned.
