In preparation of the soon-to-be-published BTBS Preview of Illinois (and all the other BTBS posts coming this summer and fall), I thought I should start by laying out exactly what goes into a BTBS post. We have a lot of new readers who may not have been around last fall, so let's start with a primer.
In early 2007, I began entering college football play-by-play data, at first just for the Big 12. I wanted to create a way to get more in-depth, telling and descriptive stats to describe the Mizzou games I had watched and would watch in the future. I perused Football Outsiders, then and now the premiere site for advanced football statistics. I stole a few measures to get myself started, and then I set out to create some of my own. It started with posts like this at Mizzou Sanity and soon got position-specific. As the 2007 season began, I used some of the stats I was playing with to put together "Beyond the Box Score" reviews of games...and then previews as well. Then came the move to SB Nation, and my nerddom grew. By the Cotton Bowl (Preview, Review), I had a pretty good system going.
Now to 2008, where things took off. I got a gig at Football Outsiders writing a weekly column called Varsity Numbers, adding (I hope) depth to their college lineup and opening up a lot of possibilities. Not many people are actually obsessed enough to actually enter 800 college games a year into a database, let alone analyze the data, so just by being persistent in my nerdiness, I had a leg up. But as time has passed, more useful measures have come about. As we embark on what will be my third season of previewing and reviewing Mizzou games, I'm confident in the direction the measures are heading.
So without further adieu, let's take a look at some of the BTBS numbers to which you'll be exposed constantly over the next seven months.
There are basically four levels at which to look at college football statistics: the year level, the game level, the possession level, and the play-by-play level.
The year level is most often just looked at in terms of wins and losses. Team A went from 7-5 in Year A to 5-7 in Year B and 3-9 in Year C; consequently, Coach A was fired. There isn't much, statistically, in this.
The game level is extremely common. Team A is averaging 37.3 points per game. Quarterback B is averaging 298.7 passing yards per game. You can rank teams and players with this, but much of the underlying story is still left untold.
The next level is possession data—at Football Outsiders, this level of analysis covered by Brian Fremeau's Fremeau Efficiency Index (FEI). It is the basis for FO's college projections (I'm still playing catch-up in that regard).
That leaves the play-by-play level, which is my territory. Measures like Success Rates (adjusted for college data), Points Per Play (PPP), and Success + PPP (S&P). The possibilities in this level of data are strong. To illustrate that, let’s go back in time.
November 10, 2007. College Park, Maryland. Tight end Jason Goode catches two touchdown passes (of ten and seven yards) from Chris Turner as Maryland upsets No. 8 Boston College, 42-35. At the "game" level, this one is easy: Maryland won, Goode had two touchdowns. He gets the "Goode: 2 TDs" treatment on the ESPN ticker for the next 24 hours. Good for Mr. Goode.
However, what contributed more to Maryland's touchdowns -- Goode's two receptions or the 43-yard catch by Darrius Hayward-Bey that set up the first score and the 45-yard catch by Isaiah Williams that set up the second? Do the higher-level college football statistics (i.e. the "year" or "game" level) give credit where it is truly due? At the play-by-play level, one can address these questions.
There are two main measures I use to evaluate play-level college data: Success Rates and Points Per Play.
Like On-Base % in baseball, Success Rates measure efficiency. To explain, let's first define what makes a play a "success." A "successful" play is determined as follows:
* 1st Down success = 50 percent of necessary yardage. If it's 1st-and-10, you need 5 yards for 'success'.
* 2nd Down success = 70 percent of necessary yardage (rounded up to the nearest yard, of course). If it's 2nd-and-10, you need 7 yards for 'success'. 2nd-and-15? 11 yards. This makes sense, really, because to succeed regularly on 3rd downs, you need to stay at about 3rd-and-5 or less. Getting most of the way there on 2nd downs sets you up infinitely better for 3rd down.
* 3rd/4th Down success = 100 percent of necessary yardage. Self-explanatory.
If a team racks up a whole bunch of yards but only has a success rate of, say, 35%, they probably are going to struggle to win the game unless their defense is great. Racking up a ton of yards in 3-4 plays while forcing yourself into 3rd-and-8's the rest of the game is not a good recipe for scoring a lot of points.
There is no better measure of efficiency than success rates. But what about explosiveness?
Points Per Play
Clearly the standard measure for explosiveness, Yards Per Play, is a good one. But all yards are not created equal. A 10-yard gain from your 15-yard line to your 25 is not the same as one from your opponents’ 10-yard line to their end zone, or one from your opponent’s 40 to their 30, advancing into field goal position. That is the thought behind EqPts, an equivalent points measure that assigns a point value to every yard line on the field based on the average number of points a team can expect to score from there. By simply subtracting the value of the resulting yard line from the initial yard line of a given play, you can assign every play an equivalent point value. This assigns credit to the yards that are most associated with scoring points, the end goal in any possession, and lessens the credit given to yards that are less useful*
* Clearly all yards are useful. Every yard you get moves the other team further from your own goal line, but still...some yards are worth more than others, and credit should be assigned with that in mind.
With EqPts, you can break a game down and build it back up again through point values. Add an offense's EqPts to the value of the penalties, turnovers, and special teams events of the game, and you get a pretty accurate look at how the game should have gone down given average luck for both teams. You also get a nice measure called Points Per Play (PPP), which takes the Yards Per Play measure and ties it more directly to points scored. This is a strong measure of a team’s, or player’s, explosiveness.
In baseball, one of the more common and useful stats in baseball is OPS (On-base percentage Plus Slugging average). It measures both efficiency/consistency and power/explosiveness. Well, if Success Rates accurately measure efficiency, and PPP accurately measures explosiveness, then why couldn’t you combine the two to come up with a number as telling as OPS? You can. The S&P measure (Success Rates plus Points Per Play) does just that.
On an individual player basis, S&P can show flaws that simpler figures like Total Yards or Yards Per Carry/Catch would not. If Running Back A is a big-play threat but has all-or-nothing success—either a huge play or a loss of yards—then his yards per carry (or PPP) will likely be pretty good. But their success rates probably will not.
On the other hand, if Running Back B gains between four and six yards every single play, he will likely have a strong Yards Per Carry figure in the 5.0 range, and their success rates will be solid as well; but their overall PPP figure will probably reveal a flaw—an offense needs both efficiency and explosiveness to succeed (actually, PPP is more closely correlated to win percentages than success rates, suggesting explosiveness is actually a hair more important), so while Running Back B would be an asset to any offense, if he were the team’s #1 weapon, the offense probably wouldn’t be very good.
S&P+ (and other "+" numbers)
The most frustrating thing about college football stats is that while one team is putting up good numbers against a good team, some other team is putting up great numbers against a terrible team. It's impossible to get too much information from statistical rankings because of this.
On August 30, 2008, Graham Harrell threw for 536 yards against Eastern Washington; meanwhile, Sam Bradford threw for 395 yards against Cincinnati on September 6. In 2007, Colt Brennan threw for 416 yards and 6 touchdowns against Northern Colorado on September 1, while Tim Tebow threw for 304 yards and 2 touchdowns against South Carolina on November 10. By all basic statistical accounts, Harrell's and Brennan's stats were insanely good and looked better than Bradford's and Tebow's performances on the ESPN scroll. However, could Brennan have put up Tebow's numbers against SC? What would Bradford have done against Eastern Washington? Thanks to the "+" concept, we can start to answer those questions, and once again begin to better assess the quality of a given team or player.
As the S&P is to OPS, the "+" number is to ERA+ or OPS+ (also known as Adjusted ERA or Adjusted OPS) figures. The goal of the "+" is to adjust for what is expected against different opponents. The goal of the "+" figures is to simply compare a team’s output versus what would be expected given the opponent. Like OPS+, a "+" score of 100 means the output was average, exactly what would be expected against the given opponents. Anything over 100 is above average, anything under is below average.
If Team A posts 25.0 EqPts in the passing game, that is a decent total; but did it come against a defense that typically gave up 30.0 EqPts in the passing game (in which case their PPP+ would be 83.3)? 15.0 (PPP+: 200.0)? The "+" formula takes into account a team’s production, the quality of the opponent, and the quality of the opponent’s opponent. It can be used to measure everything from a team’s overall S&P+ to 3rd Down Rushing Success Rate+.
Here are some of the measures you'll see often:
S&P+ - The overall S&P+ ranking combines a team’s overall offensive and defensive S&P+ for all plays that took place while the game was defined as "close" (within 24 points in the first quarter, 21 in the second quarter, and 16—i.e. two possessions—in the second half). Combining a team's offensive and defensive S&P+ gives me my overall S&P+ rankings, which I shared in the past.
Rushing S&P+ - A close-game S&P+ used only for running plays. Note: sacks count as pass plays. Standard college statistics tie sacks into rushing totals. These do not.
Passing S&P+ - A close-game S&P+ used only for pass attempts (passes and sacks).
Standard Downs S&P+ - An S&P+ figure that looks at all plays that took place in non-passing down situations. Passing Downs are defined as follows: Second-and-8 or more, third-and-5 or more, fourth-and-5 or more. Anything less than that—any first down, second-and-7 or less, third-and-4 or less, fourth-and-4 or less—are considered Standard Downs* because running and passing are more-or-less equal options. These divisions in downs were made based on raw S&P data that showed a clear division in success rates and points per play success between Passing Downs (worse stats) and Standard Downs (better). The better a team can stay out of Passing Downs, the better they will do overall, therefore looking at Standard Downs S&P+ is quite telling.
* These used to be called "Non-Passing Downs."
Passing Downs S&P+ - An S&P+ figure looks at all plays that took place in passing down situations.
Other Fun Stats
Field Position % -- A crude attempt at measuring the field position battle, "Field Position %" simply measures the percentage of a team's plays that took place in their opponent's field position. The higher number, the better...obviously. Anything over 40% is decent; anything over 50% is quite strong.
Leverage % -- This is derived from the idea that leveraging a team into as many Passing Down situations as possible is a subtle, effective way of winning games. Leverage % = Total plays run on Non-Passing Downs / Total plays. The higher number, the better. Anything over about 75 percent means the offense stayed in comfortable situations most of the game.