Assessing Pivotal Plays in Cowboys-Giants Week 14 Matchup
One of my favorite things to do when I watch the Cowboys’ games at home is track the team’s win probability throughout the night using Advanced NFL Stats‘ win probability charts. Using a model which takes the down-and-distance, score, and time into account, ANS is able to determine the probability of a team winning a game at any point in time. This information doesn’t stem from estimates, but rather years of NFL data.
It is always fascinating to see how certain plays can influence a team’s chances of winning. Punts, for example, often result in a fairly significant drop in win probability because giving away possession is generally detrimental to a team. Near the beginning of games, it takes a huge play to swing win probability in a major way. A 4th and Goal defensive stop while up six points with 45 minutes to play might result in a big bump in win probability, but that same play would be much larger–perhaps from around 50% to 100%–if the play was the final one of the game.
Using the graph from Sunday night’s game, I thought it would be fun to take a look at which plays affected the Cowboys’ win probability most significantly. Below, you can see the chart, along with 10 plays (or short sequences) which I have labeled as the most important. . .
Play 1: 64-yard pass to Hakeem Nicks on 3rd and 7 at NYG 32; 53:44 to play
- Cowboys’ WP (win probability) drops from 42 to 25 (-17%)
A 17% drop in win probability in the middle of the first quarter is a big one. This wouldn’t have been much higher even if Nicks scored. Poor coverage by Alan Ball.
Play 2: 26-yard gain by Felix Jones on 1st and 10 at NYG 42; 48:19 to play
- Cowboys’ WP jumps from 42 to 52 (+10%)
The touchdown pass to John Phillips put Dallas on top, but Jones and a subsequent defensive holding penalty put the ‘Boys in position to score.
Play 3: Felix Jones fumbles on 1st and 10 and ball recovered by NYG at DAL 14; 31:38 to play
- Cowboys’ WP drops from 57 to 38 (-19%)
This had the potential to be devastating to Dallas but they made the best of the situation by holding New York to a field goal and kicking one of their own before halftime.
Play 4: 47-yard touchdown pass to Mario Manningham on 3rd and 5 at DAL 47; 19:39 to play
- Cowboys’ WP drops from 70 to 45 (-25%)
Yielding a 47-yard score on a crucial 3rd down due to a broken coverage is heartbreaking.
Play 5: 74-yard pass to Laurent Robinson on 3rd and 10 at DAL 20; 13:17 to play
- Cowboys’ WP jumps from 38 to 66 (+28%)
Other than the blocked field goal to end the game, this pass to Robinson was the most important one of the contest for Dallas.
Play 6: 15-yard completion to Mario Manningham on 4th and 3 at DAL 37; 8:24 to play
- Cowboys’ WP drops from 78 to 56 (-22%)
Here, you can see how game situation affects win probability. The pass was only 15 yards long, but it came on a crucial 4th and 3 with just over eight minutes left to play.
Play 7: Sean Lee interception on 3rd and 9 at DAL 21; 6:50 to play
- Cowboys’ WP jumps from 62 to 89 (+27%)
I actually thought this would be more valuable to Dallas, but the fact that it came on a difficult 3rd and 9 (when New York’s chances of converting were low) likely affected the jump in WP.
Play 8: Cowboys’ three-and-out; 2:20 to play
- Cowboys’ WP drops from 88 to 67 (-21%)
Romo’s infamous incompletion to Austin hurt Dallas in a big way. If you assume Romo hits that pass 90% of the time and Dallas’ wins 99% of games following a completion, the actual dip in WP would be closer to -31%.
Play 9: Holding on Abram Elam and 18-yard completion to Jake Ballard on 1st and 10 at DAL 19; 1:21 to play
- Cowboys’ WP drops from 49 to 27 (-22%)
The holding penalty on Elam has been overlooked. DeMarcus Ware’s offside penalty was also costly, but the full extent of it isn’t factored into the WP chart because the errant snap and loss by the Giants isn’t reflected in the play-by-play.
Play 10: Blocked FG; 0:06 to play
- Cowboys’ WP drops from 44 to <1 (-43%)
And the Cowboys’ playoff chances drop from potentially around 90% with a win to now around 40% .