Panthers at Seahawks Win Probability

by John Morgan on Dec 6, 2010 12:31 PM PST

Haven't posted a win probability graph in a few weeks because of server malfunction, but yesterday's win was defined by two big swings in win probability. Want to understand how Seattle was able to take over in the third quarter and run away with it in the fourth? Check out those two double black diamonds created by Lofa Tatupu's interception return touchdown and Leon Washington's 84 yard punt return to put Seattle on Carolina's one.

Wptats_medium

Tatupu's interception flipped the course of the game.

His tackle numbers are down, but some of that is a product of a scheme that funnels tackles to the weak side. Tatupu contributes to tackles made by David Hawthorne by cutting down the lead blocker and penetrating and redirecting runs. His tackle numbers are also down because he has been slowed by a knee injury. Even accounting for all that, 2010 has been his best season since 2007. He is no longer a Pro Bowl caliber middle linebacker but he's a steady presence in coverage and his pick six against the Panthers was a Golden Age Tatupu play: Tats jumped the route and was clear to the end zone. He may not be 100%, but Tatupu does look quicker and more agile than in seasons past, and maybe that's a function of Chris Carlisle's different approach to strength and conditioning.

In retrospect, Tatupu probably deserves the game ball. It put Seattle ahead and following Washington's near touchdown return, the Panthers were forced to pass. Those two plays won the game for the Seahawks.

Tweet Comment 11 comments | 0 recs |

Do you like this story?

Comments

Pick 6s always make a big impact, of course, but is the WP impact increased

in a game situation like this where basically nothing is happening except one team’s ability to consistently run the ball (partly because some of the successful running plays are in relativity negative WP plays)?

So for example, if a pick 6 occurs in a game where both offenses put up 24 points or more, is the impact lessened? This may be so self-apparent that I ought to know that intuitively, but it’s something that hadn’t occurred to me before.

by jacobstevens on Dec 6, 2010 12:39 PM PST reply actions

WP is based on point differential, location of the ball and possession

It doesn’t take into account previous offensive/defensive production by either team (aside from point differential), so a 35-42 game looks the same as a 0-7 game in terms of WP.

by Greetings from the Lord Humongous! on Dec 6, 2010 12:43 PM PST up reply actions

So Lofa's pick 6

took WP from .69 favoring Carolina to .38 favoring Seattle, a 31 point swing.

Peyton’s first pick 6 given to Dallas took WP from .80 favoring Dallas to .93 favoring Dallas, a 13 point swing.

So I think my question is answered. I might not have asked it very clearly, but what I am seeing is Lofa’s pick 6 made such a huge impact in part because very few other plays made very significant impact in this game.

by jacobstevens on Dec 6, 2010 2:06 PM PST up reply actions

Eh, no. That would be an incorrect conclusion

Manning’s pick six happened in the second quarter of a 10-0 game. That is why it was worth 13%. After the pick six, the game was still in the second quarter, still led by the same team, but became a three score instead of a two score game.

Clausen’s pick six happened in the third quarter of a 14-10 game. That is why it was worth 31%. After the pick six, the game was still in the third quarter, but a new team was in the lead, and because less time was on the clock, that lead was more likely to be decisive.

The number of plays that “made a very significant impact” is immaterial.

Another way to look at it is, in a 21-20 game, with 10 seconds remaining, and the 20 point team on the 21 point team’s 15 yard line, a pick six would probably be worth 95%. The composition of the rest of the game doesn’t matter. What matters is that the trailing team was in field goal range and the interception return for a touchdown ends the chance of a field goal, leaves little or no time left on the clock and forces the 20 point team to score a touchdown.

Whereas a pick six accomplished by a team trailing 63-0 in the fourth quarter is not likely to be worth any win probability, because 63-7 is no more likely to be overcome than 63-0.

by John Morgan on Dec 6, 2010 2:32 PM PST up reply actions

Yeah, I forgot to mention the "time remaining"

which is pretty important.

by Greetings from the Lord Humongous! on Dec 6, 2010 2:41 PM PST up reply actions

OK thanks.

What I didn’t realize is that the WP won’t change according to how the rest of the game unfolds. What I was thinking about was the fact that there was a lot of game left to unfold, and I could see how that factored. It still does in the way that I understood, just that that does not dynamically change WP for plays as the game situation changes. It instead remains static because time remaining is already factored in in the first place.

by jacobstevens on Dec 6, 2010 2:44 PM PST up reply actions

WP basically gives everyone the same chance to go the same distance down the field

Both Tom Brady and Max Hall are given the same odds to put points on the board for a given down/distance/time remaining/score differential based on historical averages for all NFL teams over the last few years.

by Greetings from the Lord Humongous! on Dec 6, 2010 3:05 PM PST up reply actions

Here is the WP calculator

if you want to play around with different scenarios, etc:

http://wp.advancednflstats.com/winprobcalc1.php

by Greetings from the Lord Humongous! on Dec 6, 2010 12:46 PM PST up reply actions

Well said

The interception return was the thrilling moment at Qwest yesterday. Tatupu sure looked slow running to the end zone though. His straight line speed has never been great, but he had better quickness earlier in his career. Here’s hoping that he can continue to channel the hyper awareness of Ray Lewis that gives him the extra step.

by 12thman on Dec 6, 2010 12:42 PM PST reply actions