Four score and seven years ago… well, seven weeks to be exact… four scores feels accurate though. This team once had a quarterback who could do it all. Now?
The Dallas Cowboys offense has fallen on hard times and we hold these truths to be self evident. After averaging 32.6 points a game through the first five contests of 2020, they didn’t score 32 points total over the next three. In fact they’ve scored just seven touchdowns in the last six games. The faucet has been turned off, the river has run dry.
The primary difference, obviously, is the man who has lined up under center for the last two months. Dak Prescott’s ankle injury in the third quarter against New York shut all hopes for the season down. The Cowboys weren’t world beaters — that win improved them to 2-3 on the year — but they were able to score at will when they weren’t turning the ball over. Now? The offense is a shell of itself and we’re not talking about one of those impressive leatherback sea turtle shells.
With Andy Dalton and Ben DiNucci, the offense had been miserable. Putting Garrett Gilbert in, which I advocated for, made a huge difference against the Steelers, but they still managed just 19 points for the entire game. Dalton returned, and with two weeks to prepare and inside intel on the Vikings the club bubbled up for a week, but went right back in the can against Washington. The shuffle of forever-more backups definitely leads one to miss one Rayne Dakota Prescott.
In that vein, and because there’s a pseudo-bye week happening, it’s time to reminisce back to when the Cowboys had a high-powered offense. Remember when it was expected that the quarterback would have a big IMPACT on games?
Here, we’ll be breaking out a hybrid of two advanced stats, Expected Points Added and Win Probability Added, and the term we’ve settled on are Impact Plays.
EPA measures how likely a play will lead to points. WPA measures how much a play increases a team’s chance at winning the game. I feel that combining the two provides a true glimpse into what the layman would call clutchness
We’ll provide some glossary definitions and notes and observations at the end, but let’s get to the plays and highlights. All data is compiled using NFLfastR.