A Better Box Score: Simple Ways to Improve the Basic Game Recap

Box scores are intended to give a snapshot of games, and, for the most part, they work. If I look at the box score for West Virginia’s 70 to 63 victory over Baylor, I have a pretty good inkling of what kind of game it was; similarly if I look at the box score of Auburn’s infamous 3-2 win over Mississippi State. But there’s an awful lot I can’t tell from those box scores. Mississippi State had 38 rushing yards, but was that because their running game was stoned or because they took too many sacks? West Virginia and Baylor combined for over 1,000 yards passing, but were those short passes receivers took the distance or long ones down field?

Run or pass? Or neither?

Run or pass? Or neither?

I’m against overwhelming the classic box score with a variety of so-called advanced statistics. I’m a fan of these and I think they are great, but the box score is not the place for unpronounceable acronyms. So below is a non-exhaustive list of very basic, very simple, hopefully very clear changes I think would greatly improve the quality of the traditional box score.

  • Sack Yards: This is an easy one, and is unique only to college rather than the NFL, but there’s no reason that sacks should count against rushing yards. It makes quarterback rushing yards extremely difficult to decipher, especially in the age of the dual-threat quarterback, and often makes passing look more productive than it is in reality. (It also penalizes quarterbacks who do the smart thing, and throw the ball away instead of taking sacks.)
  • Tackles for Loss: Sack yards should come out of passing but all box scores should also have a simple table of negative plays as its own stand alone category. You can tell a lot about offensive and defensive styles based on the number of negative plays.
  • Completions Behind the Line: Bubble screens, rocket screens, now screens, touch passes and swing passes are an increasingly large part of offenses, and, given that these plays are nominally forward passes but are typically “packaged” with running plays, they really should be their own quasi-run/pass category. (In college and high school in particular, the rules for linemen downfield are entirely different depending on whether the pass is completed behind or past the line of scrimmage, thus further arguing for different treatment.) Call it the Percy Harvin/Tavon Austin category of plays which “all-purpose” players typically thrive on. The other goal is to remove these plays from traditional passing categories (though I think it is fair to count incompletions against the quarterback), to make passing statistics more purely a measure of downfield passing.
  • Yards After Catch/Air Yards: Brian Burke at Advanced NFL Stats has done great work on yards after catch and the corollary statistics, “Air Yards” or the amount of a quarterback’s passing yards which come purely from how far the ball travels in flight, ignoring what the receiver does after he catches the ball. Although I still abide by the belief of Bill Walsh that accurate passing is a big part of yards after the catch, Burke’s studies of the NFL seems to indicate that Air Yards is very consistent by quarterback (i.e., better quarterbacks tend to, on the whole, get more of their yards from the ball in the air versus yards after catch) while YAC is very consistent by receiver — some gets lots of YAC and some don’t. This doesn’t seem any more difficult to track for a game than measuring yards themselves, and could really illuminate both the strategies each team employed and simply the flow of the game.
  • Average Starting Field Position: Hat tip to Matt Hinton for this one, but this simple and self-explanatory statistic would explain more outcomes than likely any of the other additions I’d proposed.


I’d love to hear of any additional changes (or arguments against these proposed changes) any of you may have in the comments in the comments.

  • notque

    I run an online football game, and we have many more stats in the box score like yards per completion, and yards per attempt. TFLs are an absolute must in understanding what has happened.

    I’m aware of air yards, but have never seen the value in it. I can add it just to see, but I doubt players care at all.

    Completions behind the line I’ve never heard of or considered, I think it could have some value.

    We do catagorize Sack Yards, but I don’t remove them from rushing totals. I could, it’s just not traditional.

  • Mr.Murder

    Sack yards to rushing totals is part of the trend to throw and emphasize its positives. Fantasy scoring is usually better on 300yd games, etc.

  • IrishBarrister

    I like the idea of including the number of sacks and sack yards in the box score. I would go one step further though and add sacks to pass attempts and subtract sack yards from passing yards (and correspondingly, yards per attempt). It is not an original idea, but I think it is more illustrative of true drop back passing efficiency.

  • Modesto Koczwara

    Can you give me any suggestions on learning much more about the game of football from a fan’s perspective? I am currently reading Take Your Eye Off The Ball which is really helpful.

  • CrisBenson

    Chris I take you don’t care for all the statistical over analysis when you wrote “so called advanced statistics” and If that is the case, I am right with you. I agree box scores are way overdue for a complete overhaul but mathematicians cannot replace an biased but informed opinion of a scout or coach. I personally believe a little biased opinion is a good thing especially when a mechanical flaw is causing an athlete to over compensate. That flaw could lead to problems that stats will not reveal. Sabermetrics in baseball is a great tool but football has way too many variables With subjective observation, evaluation, and a little statistical analysis a scout, coach can come to an confident opinion.

  • IrishBarrister

    I think you might be missing the role of statistics in aiding traditional methods of evaluating players. Those advanced statistics help guide a coach or scout in focusing in on areas or seeing things that may not have been immediately apparent.

    I will use the example of the Houston Texans running game. The style of play they run on film and the fact that they had the 8th most rushing yards in the NFL would indicate that they have a good run game. But FO has them 16th, dead middle, in terms of efficiency. Why? Because Houston got stuffed at the line a lot, including in 3-and-short situations. That tells the offensive line coach where to focus: on the initial blocks at the line of scrimmage. He will probably break that general statement into more complex detail, e.g., left guard getting his helmet in the correct position on a stretch zone to the right, but the statistics help narrow his focus without losing days watching film.

    There’s only so much time in a day, and if advanced statistics can help a coach focus his time where it will do the most good quickly, I see no reason not to use them.

  • CrisBenson

    Great point IrishBarrister.

  • salt_bagel

    I’d categorize the passing game as completions, incompletions, sacks, and scrambles; all of which are broken down from a pool of “total pass plays” or “total dropbacks”. The issue then becomes how you handle the screen/swing game. I would agree that it should be stripped out of those stats and treated separately.

  • Ken Schroeder

    What about “how many plays are needed to account for half of the teams yards”. Tells you about big plays, consistency, etc.

    In a less outside the box one, I agree that passes should be broken down by dropbacks-attempts-completions-scrambles-sacks. In some order. Maybe figure out some way to not count passes travelling less than 5 yards from the line of scrimmage?

    Rushes maybe broken down by stuffs-positive yards-first down.