To mark David Tyree's recent retirement, The Fifth Down linked to Shutdown Corner's take on his signature accomplishment, "The Catch." How does that play rank with other iconic NFL plays? If only there were a way to put a number on how critical a play was toward winning a game...
Fortunately, there's WPA to give us the answer. From the win probability analysis of Super Bowl XLII:
+0.19 for Tyree’s catch
+0.41 for the TD pass to Burress.
Sadly, few will remember the 2-yd gain by Jacobs on 4th and 1 to keep the drive alive, but that play had a WPA of +.21. If Tyree doesn’t make the catch, the drive is still alive--it was ‘only’ 3rd down. If Jacobs is stuffed—that’s all she wrote.
Of course, there’s no good way to quantify the style points for Tyree’s miraculous grab or Manning’s escape from the sack.
I especially like comment #6 at the Fifth Down's post from Jay who describes the unquantifiable aspects of the play especially well. In fact, my Giants-fan friend prefers to call it 'The Play' rather than 'The Catch.' After all, the play has far more than just one player pinning the ball to his helmet.
It's too bad I can't get get WPA numbers for some of the other classic Super Bowl plays before 2000 without play-by-play data. Even if we did, the underlying model would be different due to the difference in the way the game has been played over the years. Still, if anyone wants to nominate classic plays and has the score, time, and down & distance info for the play, I can run it through the modern NFL model for an estimate.
In the recent decade, perhaps the Roethlisberger-to-Holmes 6-yd TD pass is the only comparison to the 'Catch'. But the Steelers already had at least overtime in hand, thanks to Holmes' previous 40-yard catch. It was that play that really made the difference--0.50 WPA compared to 0.38 WPA for the actual TD.
But for style, drama, narrative, and everything else, it's hard to claim the Manning-to-Tyree play doesn't take the cake.
It's too bad I can't get get WPA numbers for some of the other classic Super Bowl plays before 2000 without play-by-play data. Even if we did, the underlying model would be different due to the difference in the way the game has been played over the years. Still, if anyone wants to nominate classic plays and has the score, time, and down & distance info for the play, I can run it through the modern NFL model for an estimate.
In the recent decade, perhaps the Roethlisberger-to-Holmes 6-yd TD pass is the only comparison to the 'Catch'. But the Steelers already had at least overtime in hand, thanks to Holmes' previous 40-yard catch. It was that play that really made the difference--0.50 WPA compared to 0.38 WPA for the actual TD.
But for style, drama, narrative, and everything else, it's hard to claim the Manning-to-Tyree play doesn't take the cake.
There's always "The Catch". Down by 6, 1 minute to go, 3rd and 3 on the opponents 6. WP (according to the calculator) of .32
After "The Catch", there's 51 seconds remaining and the other team has the ball with around 0.1-0.2 WP (I don't know where the return came to, WP is .13 if a touchback, .21 if returned to the 30).
Thus, "The Catch" is worth somewhere around 0.5 WPA.
1972 divisional round
Oakland 7
Pittsburgh 6
14:38 of the 4 th quarter
Pittsburgh 4 th and 10 from their own 40
Bradshaw to Harris TD
Super Bowl XXXV comes to mind, with Kevin Dyson just missing the goal line.
Also, the play that got them out of the wild-card round, that infamous illegal forward pass on the kickoff.
Vinatieri's field goal at the end of regulation in XXXVI.
John Kasay's out-of-bounds kickoff in XXXVIII.
Maybe the (double?) reverse pass by steelers aginst seahawks? The most septacular SB-play that I've seen
Or if we also count PO, the ball that Jerome Bettis fumbled aginst the Colts, after wtich Vanderjagt missed a FG
One thing struck me after watching the play: What if Manning had been sacked insted of throwing the pass?
The Giants had 3 and 5 at the 44, worth WP 0,17. Then is Manning almost on his way down at the 38, giving them 4 and 11 with 1:12 to go, worth just 0,02 WP compared to the 0,44 in reality.
BTW: Why doesn't the WP change if Manning was sacked, or if Tyree didn't play Spiderman and catch the ball?
As a Bill fan, this hurts, but Norwood's missed kick from 47 yards probably gets made more than half the time, giving the Giants more than half a win.
actually the play-by-play for each Super Bowl can be found here: http://www.usatoday.com/sports/football/super/superbowl-plays.htm
Sweet.
What, no mention of Harrison's 100 yard interception return? Before the play the Cards had a 0.57 WP, and after the 2nd half kickoff had a 0.27 WP, for a change of 0.29 WP. Instead, if the Cards had scored a TD their WP would be 0.68 to start the 3rd.
It's not quite the same thing, but here's the All-time greatest Super Bowl players:
QB - Joe Montana 38.7
RB - Roger Craig 14.5
WR - Jerry Rice 22.8
TE - Jay Novacek 5.6
I did this analysis four years ago using Citizen Sports/Protrade's football Net Expected Scoring. I have the PBPs somewhere if you want them.
Here are the PBPs
http://www.usatoday.com/sports/football/super/superbowl-plays.htm
Nominations:
Super Bowl V: 0:59 remaining, Baltimore interception at Dallas 41, 13 yard return, in a 13-13 tie.
Super Bowl 13: 1:17 remaining, SF down 16-13, 2nd-20, Montana 27 pass to Rice from the Cin 45 to 18.
Brian,
Just a couple questions: What does CBF, EI and LI stand for in the game models (the in game fluctuating wp charts is what I mean)? Also, have you ever looked back through the games which have been modeled to see how accurate the measurements were. By that I mean you should have a decent amount of data that says at the end of the 3rd quarter x team should win y% of the time, this should then be measurable against what actually happened.
In order: CBF - Come Back Factor is a measure of how difficult a comeback was based on WP (aka score/time remaining).
EI - Excitement Factor is essentially just a measure of how exciting the game is either because the lead swung back and forth or it was close most the way (a blowout would have a very low EI).
LI - Leverage Index, but I'm not entirely sure what it is. I believe it's based on how important possession is, and is bigger later in games.
Probability Confirmation - Since all of his data is based upon actual NFL games including the score, time and eventual winner, yes all of his data is accurate. This comes up even more when he starts predicting wins based upon his per play statistics and he's checked those numbers too and they are within expected variation.
Thanks, James!
LI is indeed leverage index. I have been kicking around different versions of it, but the numbers in the WP graphs are more of a placeholder for the moment. It's the measure of the importance or 'criticality' (to make up a word) of a play. Technically, you take the WP difference between a good outcome on the play and the bad outcome on the play, and divide by the typical play's difference in WP. The trouble is deciding what is a "good" outcome (a TD pass? a 1st down? a 5-yd gain?) and a "bad outcome" (a turnover for a TD?, a sack?, a stuff?). Anyway, that's the concept.
Here is detailed data on the WP model's accuracy/calibration.