Field Gulls: An SB Nation Community

Navigation: Jump to content areas:



Around SBN: MLB Hot Stove: blogging the rumors, trades, signings Bar-right-arrows



In which I express my humility in the face of an unforecastable game

I'm not going to discount New York's title as a fluke. It certainly was improbable, whatever that's worth. It's irrational to argue that four postseason games, won by a total of 20 points, are more important than the 16 games that preceded them. No doubt New York won when it mattered most, but clutch is not an ability, it's an adjective added by writers after the fact.

Before the season started, DVOA, among the best metrics for predicting future success, forecasted New York winning 8.3 games and accumulating a 5.2%, or 12th overall DVOA. Their defense was predicted to decline-badly: 6.6% or 24th. Seattle by contrast was projected to have 9.9 wins and a 19.3% overall DVOA.

New York's title and 3-0 record (and, for that matter Seattle's ugly 1-2 record) has cast this game as a mismatch. Champion against pretender. Best team from the best conference in football against a mediocre team from the worst conference in football. Given the talent on the field, New York without Osi, Seattle just beginning to rebuild a shattered receiving corps, meaningful preseason projections and what 2008 has told us, this is a close matchup that tips in New York's favor because of home field advantage. Slightly.

In one important way, New York's defense hasn't shown its potential: forcing turnovers. Despite forcing punts on 60% of all drives, New York has forced but one turnover, an interception courtesy the hapless Rams. That will eventually revert to normal.

It's no small victory to whoop weak opponents, but they haven't really. They edged a just bad Bengals club in overtime after winning the coin toss, crushed the Rams, and beat a discombobulated Redskins outfit I'm not sure exists anymore. Again, is that to discount a 3-0 team with a +40 point differential? No, it's to say that nothing done really defies Football Outsiders first projection.

The team is off to a hot start, by any metric it dominates Seattle, and given home field advantage there's no guarantee New York won't crush Seattle into dust. It's early, the Giants could be taking the next step, becoming a perennial contender, they certainly have a good group of young talent, but is that really what I think? Is that where I'd lay my money or my life? No. I don't think the Giants are this good, and I don't buy three games as enough information to prove otherwise. Teams get to hot starts and look like world beaters, but like the 1998 Seahawks, they can crash down quickly.

In one important way, New York's offense hasn't shown its potential: allowing turnovers. One pick in 29 drives is not sustainable. Still, it's been dominant on those drives, garnering 42 yard per drive, and when it reverts one can't really know what it will revert to.

So, we're left with two teams. One with a title and a pristine record that's been a little unlucky forcing turnovers on defense, but equally as lucky limiting turnovers on offense. One with a bad rep and worse record, that's been unlucky forcing and about what you'd expect preventing turnovers. Lead ball carriers Matt Hasselbeck and Julius Jones have longstanding records of allowing minimal turnovers. The Giants are at home, a ~15.0% DVOA boost; Entering the season, Seattle was projected as a 14.1% better team.

You see where I'm going with this.

The possible outcomes of a football game run in exponential time. That is, if one attempted to reduce it to its most basic steps, and computed each possible outcome starting with the kickoff and each dependent outcome from there on, the number of basic steps it would take to project even one game would take the fastest computer on Earth many time's the history of the universe to project. When two teams are relatively evenly matched, and the data to point towards an advantage for one or the other small, then it's foolish to think an accurate prediction can be made. You really can't eliminate any possible outcome or even say one is much more likely than the other. There's good arguments based on statistics and personnel that either team could win in a blowout, a close low scoring game, a close high scoring game or a somewhat lopsided game that goes either way.

I think Seattle is a good team that's underrated. I think New York is a good team that's overrated. And I have no idea who will win or how they will win.

Tomorrow we'll talk matchups. Matchups matter, especially in close games. This game has the feel of a playoff game. Like a playoff game, the team that wins the turnover battle will likely win. Not because they're better, but because they're luckier. And that's football.

0 recs | Comment 11 comments

Read Related

Story-email Email Printer Print

Comments

Display:

Good stuff

I’m curious where you got the DVOA predictions. I’ve got PFP and I’m not seeing them, are they in the premium DB or somewhere on the FO site?

by Nate Dogg on Oct 2, 2008 2:59 PM PDT   0 recs

We need this win.

The last time we beat a team on the road that finished that same season above .500 was Denver in 2006 (they finished 9-7). Easy schedules have meant we haven’t had many opportunities but we sure did squander them when they came up (PIT, CLE, etc.). I have been thinking about it and while I think the Seahawks are quite talented and have had a lot of success in recent years, they really have not done well against good teams, especially on the road. Ignoring our excellent record against the NFC West we have been otherwise pretty middle of the pack.

Call the Giants overrated or not, they are certainly a good team and likely to make the playoffs, if not as a division winner then as a wild card team. Beating the Giants on the road is the sort of win we need right now. So our guys believe that they can.

by michaelfox99 on Oct 2, 2008 3:14 PM PDT   0 recs

Just to clarify what I said above about the unpredictability of this game...

consider the traveling salesman problem. If a salesman travels to ten cities, and can travel to those ten cities in any order, his possible routes are 10!

10 × 9 × 8 × 7 × 6 × 5 × 4 × 3 × 2 × 1 = 3,628,800

Right?

If I asked you to find the shortest route, and the cities were tightly bunched, it would be almost impossible. But if I said, what if one city is LA, the next is SLC, next Denver, next Saint Louis and so forth, the answer becomes reasonably easy because we can eliminate obvious bad routes. In fact, the above is quasi-linear: west to east.

So if Seattle were facing a terrible team, and I could say that no they won’t be very successful on offense, or no they won’t be able to stop Seattle’s rushing game, or whatever, possible game outcomes bunch and I can sort of forecast likely results. I don’t see anything like that against New York.

by John Morgan on Oct 2, 2008 3:55 PM PDT   0 recs

Good read on the game

Which is that there really isn’t one. I completely agree that it could go either way. I wouldn’t be surprised a blowout win by NY, a squeaker win by NY, or a win by Seattle, even a blowout.

What I am looking for more than a win this Sunday is how the Hawks respond in a very hostile environment on the road. They flunked in Buffalo. One mistake is ok, but two is shame on you. A good sign of things to come is if they come out of the gates faster and more focused than they did in Buffalo.

It all starts in the line play on both sides. Can we protect Matt from those nasty blitz packages? Will he be able to hit his hot reads (Engram and Branch)? How ready will those 2 key guys be for Sunday? Will our secondary shut down their WR’s? Will we be able to force some turnovers? How will our ST play?

by kmedic on Oct 2, 2008 6:03 PM PDT   0 recs

The phrase “revert to normal” doesn’t make any sense. The implication there is that it was once “normal” and it will return to that norm, which isn’t the case. The phrase you’re looking for is “regress to the mean.”

by thecassino on Oct 3, 2008 9:10 AM PDT   0 recs

It's regress towards the mean

as things don’t regress directly to the mean, but up or down towards it. Averages do not all eventually end at the mean, but ones that are well away will usually move back towards it as the sample grows.

My phrasing is just a colloquial way of saying the same thing. The phrase “regress towards the mean” is spat out so much in modern sports analysis, I’m simply looking for alternate ways of saying the same thing. I think people understood me, so I’m not so worried if it’s technically incorrect.

by John Morgan on Oct 3, 2008 12:23 PM PDT to parent up   0 recs

Comments For This Post Are Closed


User Tools

Damn, Double Bastard is tasty.

FanPosts

Community blog posts and discussion.

Recent FanPosts

Patspulpit_small
Q&A with Pats Pulpit
621hawks_calendar_37632_duckett_small
Spencer, Wahle likely headed to IR
Small
The recipe: how should the Hawks draft?
Picture_1_small
Scruffy's Video now live - Need feedback
Small
NFL Rules
Picture_1_small
Field Gulls Lounge: Eat, Drink, Socialize
Picture_1_small
Video Preview: Scruffy is out of steam
Small
Offseason Satefy Shakeup
Vp081-c_small
Post Your Hawk the Redskins
Small
Putzier Re-signed, Adams Back to Practice Squad

Post_icon New FanPost All FanPosts Carrot-mini


Managers

Image_114_small Shrug

Jj_flag_detail1_small John Morgan

Graphics Editor

Picture_1_small Scruffy Lefty

ad

Site Meter