The StatHawk, Week 6: The Seahawks and positive play differentials

Part 1: A turnover a day does not keep the victories away

First, the bad news, which is actually good news in disguise: The Seahawks are not feasting on turnovers. Through four games, Seattle has a zero turnover differential. Three giveaways, three takeaways. After watching this defense create turnovers at will last year, 2014 has been a weird experience.

Indeed, in 2013, the Hawks had already forced 13 (thirteen!) turnovers after four games. That hasn't come close to happening. And taking the ball away matters. Mediocre teams lose the takeaway battle consistently. Good teams lose when they fail to take the ball away.

The team that wins the turnover battle +1 wins 68 percent of the time; the team that wins it +2 is even more successful, prevailing in seven out of eight games. The chart below will fill you in.

It will also fill you in with the theory that great teams win even when the turnover well runs dry. See! I told you there was good news. The Hawks are not slaves to the takeaway. During the Wilson-Carroll Happy Times Era, the Hawks have scoffed at the numbers which torment other teams.

	2012 Hawks	2013 Hawks	2014 Hawks	RW+PC history	NFL 1994-2013
TO +2 or more	4-0	4-1	n/a	8-1	.875
TO +1	3-1	5-0	1-0	9-1	.680
Draw	2-2	2-2	2-0	6-4	n/a
TO -1	2-1	n/a	0-1	2-2	.320
TO -2 or more	0-1	2-0	n/a	2-1	.125

The Seahawks are 10-7 when they don't win the turnover battle. And somehow are carrying a winning record, only 4-3, but still, when they lose that battle. The rest of the league doesn't work that way.

So no, the turnovers haven't shown up yet. But they will. And when they do, the Hawks will be ready: they are 17-2 since 2012 in games where they out-steal the opponent.

But let's go worst-case scenario, and let's say that the fan never stops spinning, even after we unplug it, so in its possessed state, it keeps sending shit all over the room all season long. And let's also say that probability's middle finger gets stuck in the upright position, facing Seattle -- fumbled footballs keep bouncing back to the opponent 83 percent of the time, as they have been. Let's even say that Wagner and Thomas continue to let interceptions slip through their hands, week after week after week. Finally, just for the heck of it, let's say that the DB's do hang on to a couple ducks, only to see their apparent interceptions negated by penalties.

Even then, the Hawks won't be in a dire position. Because they still have Toxic Differential on their side.

	Standings, Toxic Differential
2013 Season (Toxic Differential = TO margin + explosive plays)	1. Seahawks (66) 2. Eagles (60) 3. Niners (42) 4. Panthers (25) 5. Broncos (17) ... 30. Cowboys (-29)
2014 Weeks 1-5 (remembering that Seattle has played only four games)	1. Niners (21) 2. Seahawks (14) 3. Bills (12) 4. Broncos (12) 5. Ravens (8) ... 8. Cowboys (5)

But how do we know that Toxic Differential means anything? Who cares if the Niners are +42 in some random stat if that doesn't translate to wins? Well, there's always this:

wDVOA, 2013

1. Seattle

2. Denver

3. Philadelphia

4. Carolina

6. San Francisco

Seems legit. Weighted DVOA confirms that Toxic Differential means something, at least. Also, Seattle apparently doesn't neeeeeed turnovers to climb the table, but they sure would be nice.

Part 2: GDBS

For this second section, I'm going to trot out a toy stat I've been fiddling with. Its lame name is

"GoodDouble/BadSingle"

and it measures a quarterback's impact on a game or a season, through his most crucial plays.

(You can call it GDBS for short, which definitely does not stand for two compound curse words slammed together.)

As befits any shallow stat, GBDS is simple to calculate. You go:

Touchdowns x 2 Minus    (Interceptions + Fumbles)

That's it. The rationale is that it tallies a QB's most impactful plays only. His touchdowns produced (passing or rushing) are huge positives, so they count double; his giveaways or attempted giveaways are smaller negatives, not always resulting in opponents scoring, so they count single. The thinking is that a quarterback who puts touchdowns on the board is better than one who doesn't; but scoring six points is way more good than being careless with the ball is bad. (Feel free to re-read that last sentence, I know I did.)

A Tom Brady, grand cru 2007, would get an 86 GBDS; last year's vintage of Andrew Luck nets a commendable score of 31; but the 5-liter box of 2011 Tarvaris Jackson checks in with a score of 5.

Fine, okay, Manning 2013 comes in at 89, the highest score so far within my research. Whatever, nice job Peyton, yada yada yada.

Reaching 100 is basically impossible -- it would take an epic season with 53 touchdowns, three picks, and three fumbles.

The recent history of the Hawks and Cowboys, as told through GBDS, confirms existing conventional wisdom: the Seattle defense is hard on quarterbacks, while Dallas' is not.

	Team record	GDBS	Notes!
2013 Wilson	13-3	54-21 = 33	12 fumbles doesn't help
2013 Romo	8-8	62-24 = 38	Only a 1.9 INT% by the way
...	...	...	...
2013 Sea opp.	3-13	-5	Good
2013 Dal opp.	8-8	47	Bad
...	...	...	...
2014 Wilson	3-1	18-5 = 13
2014 Romo	4-1	19-6 = 12	So similar
...	...	...	...
2014 Sea opp.	1-3	10	Rodgers Rivers Manning
2014 Dal opp.	1-4	10	No score higher than 4, yet

• By GDBS, Romo actually outproduced Wilson last year. That's understandable once you remember RW committed a league-leading 12 fumbles. Maybe he needs bigger hands. This year shows the two QBs neck and neck again, though Romo's played an extra game.
• Turns out these guys protect the ball each about as well as the other. Perception isn't everything.
• Romo's Cowboys suffer from a poor defensive history, however. Four quarterbacks posted a single-game GBS score of 7 or more against Dallas last year: Peyton Manning with 7 in Week 5, Drew Brees with 8 in Week 10, Jay Cutler with 8 again in Week 14, and Aaron Rodgers put up a pretty 7 one week later. Good quarterbacks can beat , and have beaten, this Cowboys defense recently.
• Meanwhile, no QB has cracked 7 against Seattle since 2012.
• GDBS differential is where the stat starts to mean something, I believe. The Hawks carried a plus-38 last year in that category, while the Cowboys ended up at minus-9 -- despite Romo's tendency to make for more positive plays than negative ones.
• This year, the Boys are running a positive GDBS differential (+2), as are the Hawks (+3). Interesting to see if that holds after tomorrow's game.

Conclusion: Add it all up

Combining recent Toxic Differentials with the quarterbacks' GDBS differentials, and isolating the results to the two teams meeting on Sunday, will offer us one clear conclusion: the Seahawks are far less likely to turn the ball over and far more likely to post explosive plays. Which feels right. Here's to it feeling so right on Sunday evening.

I've prepared exactly those numbers. Weighting the 2014 figures twice as heavily as the 2013 figures (because recent results count more), and adding them all together, ought to give us a good idea of how big plays are going to shake out tomorrow. The higher the number, the better the team's outlook. So, presented without comment:

	Hawks	Cowboys
2013 Tox Diff. x 1	66	-29
2014 Tox Diff. x 2	28	10
2013 GDBS Diff. x 2	76	-18
2014 GDBS Diff. x 4	12	8
Total (higher = better)	182	-29