≡ Menu

Last Thursday, the Lions hosted the Packers, and jumped out to a 17-0 lead before the end of the first quarter. A field goal attempt midway through the 3rd made it 20-0, although the Packers then responded with two quick scores. With seven minutes left in the 4th, Detroit made it a 9-point game, and seemed to have this one locked up. In fact, when the clock hit triple zeroes, the Lions were ahead, 23-21.

Of course, a phantom face mask penalty meant the game was not yet over, and Green Bay won on the final play of the game. For the game, Green Bay had a Game Script of -10.1, which represents the average score in the game, from the Packers perspective, across the 3600 seconds of action. That’s the worst Game Script by a winning team since San Diego won in similarly remarkable fashion against the 49ers in week 16 of the 2014 season.

Below are the Game Scripts data for week 13:

TeamH/ROppBoxscorePFPAMarginGame ScriptPassRunP/R RatioOp_POp_ROpp_P/R Ratio
CIN@CLEBoxscore3733417.5233341.1%411968.3%
SEA@MINBoxscore3873116303645.5%321666.7%
DEN@SDGBoxscore1731412273940.9%392362.9%
PITINDBoxscore45103511.5393254.9%391868.4%
ARI@STLBoxscore2732411.2433654.4%351570%
BUFHOUBoxscore302195.3223637.9%452663.4%
TENJAXBoxscore423934.3333548.5%382065.5%
PHI@NWEBoxscore352873.2253343.1%612471.8%
MIABALBoxscore151322.7202643.5%492665.3%
TAMATLBoxscore231940.7283743.1%481872.7%
DAL@WASBoxscore19163-0.5302455.6%342557.6%
KAN@OAKBoxscore342014-0.8262254.2%522468.4%
SFO@CHIBoxscore26206-1.5362361%324243.2%
CAR@NORBoxscore41383-2.4423356%441475.9%
NYJ@NYGBoxscore23203-5.4532468.8%372460.7%
GNB@DETBoxscore27234-10.1392461.9%382560.3%
  • Cincinnati and Seattle posted monster Game Scripts numbers in blowout victories. Denver won by 14 points over San Diego, but even that is a little misleading: the Game Script was +12.0, because Denver led 14-0 before the end of the first quarter, and 17-0 in the second quarter. Neither the Broncos nor Chargers scored in the second half, a feat that has occurred just three times since the start of the 2009 season.
  • Buffalo continues its run-happy ways. The Bills were the only team this week to pass on fewer than 40% of all plays, in a relatively tight game.  Philadelphia and Miami — yes, that says Miami — were also really run-heavy in competitive games, with each team passing on less than 44% of plays. And Tampa Bay and Chicago (in a losing effort, but with a positive Game Script) had similar numbers.  After a really pass-happy week 12, a number of teams showed first commitments to the running game in competitive contests.
  • One team that was decided not interested in establishing the run: New Orleans. The Saints actually had a positive Game Script in the loss to Carolina, but finished with 44 pass attempts and just 14 runs.  The other team that stood out for its pass-happy ways last week was Atlanta: in a back-and-forth game with a neutral Game Script, the Falcons passed on 73% of plays, while Tampa Bay passed on just 43% of plays.  That says a lot about the tendencies of both teams.

What stands out to you?

  • Alex

    I think the Game Script from the Carolina-New Orleans game is a little misleading in terms of predicting play selection. New Orleans was trailing every time they possessed the ball in the second half, so it makes sense that they passed a lot.

    • Interesting point. One way to do Game Scripts is to measure the points differential at every second of the game for the team on offense, which would help account for games like NO/CAR where teams are trading touchdowns. The downside is that you lose the elegance of Game Scripts, as it no longer measures the score at every second. This means the two teams would have different (and not mirror image) Game Scripts, and that the average Game Script for the league would be negative. So I don’t really like that.

      • Alex

        The question is what’s the primary purpose of Game Scripts? If it’s to measure how competitive a game was, then I agree, you want one number for the game. But if the purpose is to provide context when analyzing play selection, as suggested by the name, then the most natural thing is to have separate numbers for each team.

        Now that you have me thinking about ways to improve Game Scripts, it also might make more sense to look at average score difference per unit of time remaining in the game on a play-by-play basis rather than average score difference per unit time, but that’s obviously a lot more work to calculate.