NOTE: Confused? Don't miss the definitions and footnotes at the bottom.
Our projections at Football Outsiders and in the Football Outsiders Almanac are based on precedent. Sweet, sweet regression equations and precedent. The weights of all the different factors are based on how well they have correlated to success in the past. Makes sense, right? As a whole, our Football Outsiders Almanac 2010 projections looked alright -- obviously we didn't see Texas' collapse coming, or Auburn's surge, but most of the picks were defensible. And besides, no model was going to see those events coming.
For the most part, we were reasonably satisfied with the projections when they came out, but a few teams immediately jumped out as "What the...?" candidates, and most of them had something in common: new coaches. Texas Tech, Tennessee and Kansas were all projected to do much better than most expected, so before the season began, I attempted to get a better read for what we could learn about coaching changes. Were there specific types of coaches who were likely to do better in specific types (i.e. BCS or non-BCS) of jobs? Were teams who had achieved at a higher or lower recent level (compared to their history) more prone to collapses or surges when coaches leave?
In all, I ran across some interesting data on the subject. And the data added almost no strong predictive value whatsoever. Texas Tech, Tennessee and Kansas still ended up projected far too high, and almost no progress was made on the topic.
The reason for data uncertainty in this regard is simple: it takes very little time to find an example of a specific type of coach who succeeded in a specific job ... and the same type of coach who failed in the same type of job. There is precedent for everything. But in general, one type of coach succeeds more often than any other at a BCS conference job: a coach who has already succeeded to some degree at a BCS conference job. And in bringing in former Connecticut head coach Randy Edsall after they pushed out Ralph Friedgen, Maryland got themselves one of those coaches. After well over a decade at UConn, Edsall agreed to come to his "dream job" in College Park to see if he could not only reattain the success Friedgen had in his opening years at UM, but maintain it.
2010 Schedule & Results*
|Record: 9-4 | Adj. Record: 9-4 | Final F/+ Rk**: 31
|Date||Opponent||Score||W-L||Adj. Score||Adj. W-L|
|6-Sep||Navy||17-14||W||29.7 - 26.8||W|
||62-3||W||23.0 - (-4.6)||W|
|18-Sep||at West Virginia||17-31||L||26.4 - 29.1||L|
|25-Sep||Florida International||42-28||W||34.1 - 29.4||W|
|2-Oct||Duke||21-16||W||13.8 - 24.3||L|
|16-Oct||at Clemson||7-31||L||28.4 - 16.7||W|
|23-Oct||at Boston College||24-21||W||22.6 - 29.2||L|
|30-Oct||Wake Forest||62-14||W||34.4 - 9.1||W|
|6-Nov||at Miami||20-26||L||17.2 - 31.5||L|
|13-Nov||at Virginia||42-23||W||38.5 - 24.1||W|
|20-Nov||Florida State||16-30||L||29.2 - 21.8||W|
|27-Nov||N.C. State||38-31||W||34.5 - 27.5||W|
|29-Dec||vs East Carolina||51-20||W||36.4 - 7.0||W|
|Points Per Game||32.2||29||22.2||38|
|Adj. Points Per Game||28.3||54||20.9||20|
Friedgen's final season at Maryland was his best in a while. After winning 31 games in his first three seasons (2001-03), Fridge's Terps had managed a winning record just twice in six seasons, but with new offensive coordinatoraboard, Maryland bounced back ... with defense. To be sure, the offense had its moments, and it produced above average (at least) performances in five of the last six games, but the Terps won three games in which they scored 24 points or less because they had a defense that ranked 22nd in Def. F/+, and it bailed them out when Franklin was unable to push the right buttons.
The Turtles actually still had a chance at a division title until their mid-November home loss to Florida State, but they wrapped up the Fridge's tenure (athletic directorannounced they were buying out the final year of his contract the week before the Military Bowl against East Carolina) with two impressive wins and their best final record since 2006. Evidently they decided that this level of play was not sustainable with Friedgen, and they made a move. I've said before that changing coaches is the most effective way of either getting ahead or falling way, way behind, and you shouldn't make a move unless you know without a doubt that the current coach isn't going to take you where you want to go. With just three winning seasons in seven years, it's not impossible to see what Maryland was thinking here, but the timing was still a bit odd.
|RUSHING||77||90||56||Adj. Line Yards:|
|Standard Downs||48||71||38||Adj. Sack Rate:|
|Q1 Rk||61||1st Down Rk||29|
|Q2 Rk||31||2nd Down Rk||62|
|Q3 Rk||59||3rd Down Rk||68|
As always with coaching changes, we begin by looking at the offensive footprint of the new coach's last team:
Compared to Maryland's offense last year, UConn ran about six to ten percent more and played at a faster pace (not to be confused with a fast pace ... just faster than the terrapin-slow speed with which Maryland moved between plays). Though Maryland ran Franklin's version of the spread, they still mixed in plenty of bigger packages with fullbacks, second tight ends, and the like, just like what Edsall did at UConn. So from a personnel standpoint, this isn't exactly like Nebraska moving from Frank Solich and the Power-I to Bill Callahan and the West Coast Offense.
The main problem from a personnel standpoint is ... a lot of the more talented members of said personnel are gone. Running back Da'Rel Scott (708 yards, 5.8 per carry, +6.0 Adj. POE) ran out of eligibility, playmaking receiver Torrey Smith (1,055 yards, 15.7 per catch, 63% catch rate, 12 TD) declared for the draft and became a Baltimore Raven, and possession receiver Adrian Cannon (324 yards, 9.0 per catch, 62% catch rate) and honorable mention all-conference tackle Paul Pinegar both graduated as well. Only four official starters depart, but they were four pretty good ones.
Luckily, Edsall and coordinator Gary Crowton (yes, this Gary Crowton) have two skill position players around which to build: quarterback Danny O'Brien (2,438 yards, 7.2 per pass, 57% completion rate, 22 TD, 8 INT as a redshirt freshman) and running back Davin Meggett (720 yards, 5.7 per carry, +6.5 Adj. POE). Meggett in particular should see the biggest increase in responsibility, inheriting an offense that features a much higher percentage of runs and no longer having to share the backfield with Scott.
- Few receiving tandems were more frequently targeted than Smith and Cannon. A whopping 45% of Maryland's passes were directed at one of the two (Smith 29%, Cannon 16%), meaning that though a lot of receivers return, we know almost nothing about them.
Returning Receivers By Target Rate
Kevin Dorsey (8% target rate, 15 receptions, 187 yards, 54% catch rate, 2 TD)
Quinton McCree (7% target rate, 16 receptions, 188 yards, 67% catch rate, 1 TD)
Ronnie Tyler (7% target rate, 13 receptions, 149 yards, 54% catch rate, 1 TD)
Matt Furstenburg (tight end) (6% target rate, 12 receptions, 206 yards, 52% catch rate, 1 TD)
In Fun With Small Sample Sizes, McCree seemingly showed the most reliability, and Fursternburg could be an interesting weapon, but ... who knows?
- Typically when I see huge gaps between a line's Adj. Line Yardage rankings and their Adj. Sack Rate, I wonder about how the mobility/decision-making of the quarterback played a role. In this case, Maryland's run-blocking was quite poor, but either they were agile enough to protect O'Brien or O'Brien was just really good at evading the rush or throwing the ball quickly, because their sack rates were much better. With four returning starters on the line, there will be quite a bit of pressure to upgrade the run blocking a bit. All four returning starters finished the spring on the first-string, but there was some shuffling -- R.J. Dill moved from left tackle to right tackle, Justin Lewis moved from right guard to left tackle.
|RUSHING||28||35||37||Adj. Line Yards:|
|Standard Downs||23||26||22||Adj. Sack Rate:|
|Q1 Rk||71||1st Down Rk||21|
|Q2 Rk||15||2nd Down Rk||24|
|Q3 Rk||11||3rd Down Rk||32|
Former Southern Miss defensive coordinator Todd Bradford takes over a Maryland defense that was quite solid overall in 2010. They were above-average to good in almost every category -- they were a little leaky on passing downs, but nothing debilitating -- and spectacular in one specific way: they went after the ball. Bradford's USM defenses were great against the run and played a very aggressive style, but they were frequently burned by big plays in the process. In theory, then, the returning personnel fits Bradford's style, and compensates for weaknesses, quite well.
Like teams like N.C. State or Iowa State, last year's Terps played played with the "bend, don't break, and eventually force a turnover" philosophy, only Maryland both bent less and forced more turnovers. Of their top 16 tacklers, only two were without a forced fumble or fumble recovery. This was a strength of Edsall's defenses at UConn too, so I don't expect this to change much, especially not with Kenny Tate roaming around. Maryland has produced its share of freakish athletes in recent years, and at free safety, Tate combined some serious size (6-foot-4, 220 pounds) with an incredible stat line: 79.0 tackles, 8.5 TFL/sacks, 3 INT, 4 FF, 4 PBU). Tate moves to the "STAR" position in Maryland's alignment (basically a SLB/safety hybrid), while a pair of big underclassmen -- Matt Robinson (6-foot-3, 220; 23.5 tackles, 2 FF as a freshman) and Eric Franklin (6-foot-2, 205; 16.0 tackles, 3 INT) -- inherit the more traditional safety roles.
Tate's move toward the line of scrimmage comes for two main reasons, it appears: 1) he was already playing close to the line of scrimmage as a free safety last year, 2) he already looks like a linebacker, and 3) the Terps must replace two stellar playmakers (Alex Wujciak and Adrian Moten, who combined for 134.0 tackles, 11.0 TFl/sacks, 6 INT and 12 PBU) in the LB corps. Tate also gives the Terps some flexibility to seamlessly shift between an attacking 4-3 and a faster, more reactionary 4-2-5 from play to play.
- The Terps had no national-level strengths, but they were rather strong just about everywhere. If there was a discernible weakness, it came in the front four, but quite a few interesting players return up front. At tackle, Joe Vellano (43.5 tackles, 10.5 TFL/sacks) is a potential star, while A.J. Francis (7.0 TFL/sacks) and converted end Justin Anderson (5.5 TFL/sacks) could be solid. The end position is a little iffier, with Isaiah Ross (4.5 TFL/sacks) and converted LB David Mackall looking to make waves.
- The back seven, however, could be loaded. Linebackers Demetrius Hartsfield (60.5 tackles, 6.5 TFL/sacks) and Darin Drakeford (28.5 tackles, 5.5 TFL/sacks, 2 PBU, quite all-or-nothing) have potential, and there are a couple of really intriguing cornerbacks in Cameron Chism and Trenton Hughes, who combined for 90.0 tackles (possibly too many for cornerbacks), 4.5 TFL/sacks, 1 INT, and 17 PBU.
Maryland's 2010 Season Set to Music
Sorry, Fridge, but Maryland wanted "More Than This."
(Much more palatable version here.)
Fun Stat Nerd Tidbit
Summary and Projection Factors
Below is a small handful of projection and change factors most pertinent to the Football Outsiders' preseason projections you will find in this summer's Football Outsiders Almanac 2011.
|Four-Year F/+ Rk||50|
|Five-Year Recruiting Rk||35|
|TO Margin/Adj. TO Margin****||+15 / +13.5|
|Approx. Ret. Starters (Off. / Def.)||14 (7, 7)|
Say this much for the Edsall hire: it fits the personnel quite well. I was really intrigued by the thought of Mike Leach running the show in College Park, but there are benefits to Edsall. The thing that typically trips up teams with new coaches is the occasional delay that comes with installing your own system and scheme. Aside from running more and making quite a few position changes on defense, this should be at least a semi-smooth transition.
Because of their ability to go after the ball, Maryland produced a significant turnover advantage that might be at least somewhat sustainable. It prompted a very positive YPP margin, and while things may even out a bit, the regression to the mean might not be completely crippling. There is experience on both sides of the ball, and perhaps just as importantly, there are potential stars too -- O'Brien and Meggett are good (and could be great) on offense, while Vellano and particularly Tate are standouts on defense.
Without a doubt, there are question marks. The receiving corps is a complete mystery, as are both lines, really, but Edsall should be able to make something out of this team rather quickly. Who knows what their ceiling is, but there at least shouldn't be much of a drop-off. There better not be, at least, because the season begins with two defining home games: Miami and West Virginia. Both of those teams are undergoing transitions of their own, so who the hell knows what to expect, but those two games will create the narrative for the rest of the season. Are the Terps going to be darkhorse contenders for a division crown? Are they going to be fighting and scrapping for bowl eligibility? We'll probably know by the evening of September 17.
* For more on the 'Adj. Score' and 'Adj. Record' measures below, feel free to read this Football Outsiders column. Adj. Score is a look at how a team would have performed in a given week if playing a perfectly average team, with a somewhat average number of breaks and turnovers. The idea for the measure is simple: what if everybody in the country played exactly the same opponent every single week? Who would have done the best? It is an attempt to look at offensive and defensive consistency without getting sidetracked by easy or difficult schedules. And yes, with adjusted score you can allow a negative number of points, which is strangely satisfying.
** F/+ rankings are the official rankings for the college portion of Football Outsiders. They combine my own S&P+ rankings (based on play-by-play data) with Brian Fremeau's drives-based FEI rankings.
*** What is S&P+? Think of it as an OPS (the "On-Base Plus Slugging" baseball measure) for football. The 'S' stands for success rates, a common Football Outsiders efficiency measure that basically serves as on-base percentage. The 'P' stands for PPP+, an explosiveness measure that stands for EqPts Per Play. The "+" means it has been adjusted for the level of opponent, obviously a key to any good measure in college football. S&P+ is measured for all non-garbage time plays in a given college football game. Plays are counted within the following criteria: when the score is within 28 points in the first quarter, within 24 points in the second quarter, within 21 points in the third quarter, and within 16 points (i.e. two possession) in the fourth quarter. For more about this measure, visit the main S&P+ page at Football Outsiders.
**** Adj. TO Margin is what a team's turnover margin would have been if they had recovered exactly 50 percent of all the fumbles that occurred in their games. If there is a huge difference between TO Margin and Adj. TO Margin (in other words, if fumbles and unlucky bounces were the main source of a good/bad TO margin), that suggests that a team's luck was particularly good or bad and might even out the next season.
*****Phil Steele has long tracked Yards Per Point as a means of looking at teams that were a little too efficient or inefficient the previous season. A positive Yds/Pt Margin means a team's offense was less efficient than opponents' offenses, and to the extent that luck was involved, their luck might even out the next year.