This analysis builds off the work from Chase Stuart at Football Perspective for determining a better calculation than traditional score differential.
Chase explains it as: "The term Game Script is just shorthand for the average points differential for a team over every second of each game."
My code uses Armchair Analysis data to calculate the score differential for each team for every second of every game. That is then compared to the pass ratio to see if teams are more pass- or run-heavy than you'd expected based on the treadline for the entire league.
In this example, I plotted the New England Patriots' game scripts against the backdrop of the entire 2015 season. You can see that the Patriots are extremely pass-heavy, with only two games all season long where they ran more often than expect based on the league trendline.
I used this analysis last season to show that Lamar Miller's lack of rushing attempts with Joe Philbin at the helm was more about negative game script than ignoring the run game.