From: Michael Salfino
Date: Wed, Feb 1, 2012 at 12:49 PM
Subject: Super Bowl Breakfast
To: Scott Pianowski
What's worse for a Jets fan than a Giants-Patriots Super Bowl? How can you root for the Evil Empire? But if they lose, you have to deal with insufferable Giants fans rubbing another ring in your nose like they earned it. They are not good winners, at least not here in North Jersey.
We have a game to dissect. Let's start with Massey-Peabody's pick because I think they have focused on the central issue, one with very broad relevance when it comes to forecasting games. Does recent performance outweigh full-season performance? Let's call this base-rate vs. bits, because the evidence over many years says that full-season performance is by far the best measure. There are exceptions, but can't we always rationalize that we've found one? There's the danger.
What are you going to be looking at as keys? Here are mine:
- Is Gronkowski hurt or injured? Everyone plays hurt, but if you're injured you can't perform. That high ankle sprain may keep him from being the queen on Tom Brady's chessboard, as he's been all year. (With apologies to you and Marshall Faulk.)
- Can the Giants get inside pressure? Brady steps up very deftly to evade outside rushers but there is no counter to the inside rush. It doesn't have to be the tackles. They can stunt.
- Who covers Victor Cruz in the slot? Julian Edelman? That seems like a disaster in the making for me. The Patriots haven't faced a good slot receiver in the postseason, never mind the best in football by most objective measures.
- Do the Giants shadowbox and play with confidence that their defense is going to be stout, or do they attack aggressively with their wideouts, expecting to need 30-plus to win? What's the Giants run/pass split? Every Giants run on Sunday is a win for the NE defense, IMO.
Finally, did you watch "Namath" on HBO? Well done, I thought. If not, do so and tell me what you think.
From: scott pianowski
Date: Thu, Feb 2, 2012 at 11:09 AM
Subject: Re: Super Bowl Breakfast
To: Michael Salfino
Finding the best sample size for NFL evaluation is as easy as catching a butterfly with your eyes closed. Players are always healing and getting hurt, schemes jell and schemes fall apart, conditions and contexts change. I know this, I care a lot more about recent results than anything that happened in September - for my money, those results might as well be from the 1970s. (I haven't seen the Namath doc yet, but I'll try to get to it before we close; it's on my DVR.)
One example proves nothing, of course, but we have to at least mention in passing the setup from Super Bowl 42. The Patriots were an untouchable juggernaut into the middle of November, covering nine of their first ten games and posting a ridiculous 445 points. But before anyone could say "Mosi Tatupu" the bubble burst - the Pats scored a mortal 244 points in their last nine games, covered just once, and, of course, didn't beat the Giants in the Super Bowl. Damn you, David Tyree.
The Giants were all over the map that year. A new defensive scheme was slow to take and Big Blue allowed 80 points en route to an 0-2 start. Six straight wins followed, but the Giants were nothing special after their bye - just 4-4. A close 38-35 loss to the Patriots in Week 17 was a moral victory, but I don't know anyone who expected the Giants to go deep in the playoffs, let alone win three straight games in difficult locations (Tampa Bay, Dallas, Green Bay: underdogs at every stop).
So Super Bowl 42 presented the same sort of dilemma we face today. Were the truest Giants the tough customers we saw in the playoffs, or the ordinary gang that limped to a 4-4 finish? Where the truest Patriots the "name the score" group that pasted the league in the first half, or the good-not-great club that had trouble with Jacksonville and San Diego in the playoffs? While I certainly didn't forecast an upset for Super Sunday, I do know this: I thought the 12.5-point spread was a joke, and the Giants were the obvious ATS value. I took the conference playoff results seriously, not to mention the close game the teams played in Week 17. After the game, I made the case that it was more of a perception upset than a fluke upset: the Giants were a lot closer to the Patriots than the public wanted to accept.
Mind you, there's an odd feel to both of these teams, an unsatisfying residue from Championship Week. Obviously the Patriots were damn lucky to escape against Baltimore, and while it hasn't been harped on as much, the Giants can say the same thing for their overtime victory at San Francisco. There were other lucky elements at play: the Pats were fortunate that Denver somehow beat Pittsburgh (the Steelers are a horrible matchup for New England), while New York had that Hail Mary at the end of the first half at Green Bay. (The Giants were the better team anyway that day, but it's still a case of sliding doors if Hakeem Nicks doesn't make that play - we'll never know what winds up being different.)
But hey, I'm just glad we're not forced into a Joe Flacco-Alex Smith filibuster this week. This is the game we want, the buzzy quarterbacks, the marquee head coaches. (Flacco tried to pull off the Fu Manchu and failed, miserably. Even smart phones have an inside joke with Flacco: look what his last name auto-corrects to.)
The two-week break certainly helps the Patriots: their second-most important player (Rob Gronkowski) gets extra time to heal, and obviously Bill Belichick is aided by the extra study time. Belichick's defensive game plan is the ticklish part of the assignment. How on earth do you cover New York's outstanding receivers (Nicks and Victor Cruz are stars, Mario Manningham a pretty damn good No. 3) with New England's scrubby secondary? Belichick needs to find a way to hit Manning or make him uncomfortable, but do the Patriots have the personnel?
Manning actually struggled against the nickel and dime packages in the first meeting, though Nicks didn't play that day (Ahmad Bradshaw was also out). Now he's got all of his pieces in place, and Cruz has a full season of experience under his belt. This offense is scary.
I'm not nearly as concerned with New England's offense. Tom Brady rarely plays two poor games in a row. So long as Gronkowski isn't hobbling around like Fred G. Sanford, they'll find a way to make him useful - even if it's as a decoy. They do more shifting and cute stuff with Aaron Hernandez anyway; generally it's Hernandez they scheme with, while they simply tell Gronkowski "hey, go run over someone." Gronkowski's also a much better blocker.
It's a shame the Patriots don't have a solid No. 2 wideout, someone to threaten the intermediate and deep parts of the field. But I still feel confident New England gets to 27 or more points with relative ease.
Will it be enough? I'm still mulling it over. Your witness.
From: Michael Salfino
Date: Thu, Feb 2, 2012 at 12:48 PM
Subject: Re: Super Bowl Breakfast
To: scott pianowski
Of course, I disagree that finding the best sample size is hard. M-P found it - it's the full season (though recent game are weighted slightly more than earlier games in line with their statistical findings). For me, the injuries - generally - tend to even out. Schemes gelling and falling apart may just be a convention of our minds in trying to find a pattern for normal variance relative to the baseline (baseline being the full-season rankings, especially at this stage). What's hard is avoiding temptation to use the most recent and thus most available sample of data (to our minds) to make definitive judgments. I stipulate, of course, that some teams definitely get significantly better or worse. But I don't see how we can tell who those teams are. Look at the 2007 Giants: If Tyree doesn't catch a ball off his head, there goes the whole "identity axis" argument (queue Liss). Then it's the Giants got lucky by having a receiver (wasn't it Miles Austin?) drop a certain touchdown pass and then had Favre basically give a game away, and who cares about beating Jeff Garcia?
For me, the Giants winning these games isn't shocking. But I am shocked by the performance of their secondary. That YPA number (going back to the Jets game) is ridiculous, and it's against some good-to-great QBs. I know that the Packers dropped a lot of passes, and Rodgers missed a wide open Jennings for an early, 40-yard-ish TD, but still...
Speaking of Rodgers, the most interesting thing for those who don't know Namath's history is the Alabama highlights before his knee injury when he looked basically like Aaron Rodgers on steroids. Plus speed and running ability. Maybe plus-plus. And of course, that arm. So what we saw with Namath was - at best - on one leg (and often on none, really). They also did a great job on that Jets-Niners game from 1971 when Namath came off the bench after not playing for about two years. Here's my definition of a Hall of Famer - when the home fans give you a 15-minute, post-game standing ovation after a game you LOSE, you are a Hall of Famer.
Bradshaw has a broken foot, so how much he's in is a mystery. I can't see Gronkowski being effective. I disagree with Hernandez - he's for show but Gronkowski is for dough - a mismatch against everyone because of his size. Hernandez is short.
The thing that gets me about this game is that every Giants fan of course thinks there is no doubt whatsoever that the Giants win. But also all of the national experts are picking the Giants. Many think the Giants should be favored - a team minus-6 in point differential versus plus-171 for the Patriots. That's the fourth-biggest differential ever, and the 11 games with the biggest had an average line of about 11.5 points. Am I really saying the Patriots are getting no respect? Well...
I see Brady coming out with fire in his eyes. Remember, he put up 20 points in the second half against the Giants last time and scored the last four times he had the ball. The Giants shadow box too much and screw around with their crappy running game - 18th in first-half points in the regular season, and they trailed in all their playoff games. And if this is their defensive priority, Giants fans should be face palming right now:
Our whole principal is to stop the run first, more so than stop the pass. If we make them one dimensional it's an easier game for us.
- Giants defensive coordinator Perry Fewell
WHAT?!? So if the Giants were told that Brady is going to throw on every play, they'd have no worries? I can't believe how coaches cling to demonstrably false notions about a game they spend their life toiling in. On a related note, if the Giants lose this game 37-27, how soon before the "experts" blame it on their last-ranked running game and cite improvements there as a 2012 priority?
My pick: Gronkowski is a huge issue. If he were healthy, I'd be a lot more confident about this pick. But I think the Patriots will get pressure on Eli (they have about double the Giants post-season sack rate, and Big Blue beat them 8.2 to 6.5% in the regular season.) I think the Giants fall behind early and then claw their way back before Brady finishes them off in the fourth quarter. I don't see the Giants wining the turnover battle, which has keyed their last two victories. So Patriots 30, Giants 23.
From: scott pianowski
Date: Thu, Feb 2, 2012 at 11:24 PM
Subject: whitey ford, emerson boozer
To: Michael Salfino
Here's the crux of it - we agree that teams will get better or worse as the season goes along. But we also see that it's very hard to measure that improvement, to the point that the full-season measure is better (more predictive) in the majority of cases. I can accept that as a ground rule for the league as a whole, but that doesn't mean the general rules apply to every example, or we should try to stop trying to figure out which way the wind is blowing with the teams in front of us.
If you fail to prove something exists, you haven't necessarily proven that it doesn't exist. Maybe in five years, or 10 years, or 23 years, we'll have measurements that show us what improvements look legit and what improvements look false. Statistical analysis is always evolving, and football got a late start compared to other sports. I refuse to look at any of this as a open-and-shut case, something that's even close to definitively resolved.
At least we all agree that the recent games need to carry more weight.
I watched Namath today, and it indeed was terrific. (HBO is the gem of cable, the most irreplaceable channel.) It's impossible to come away from that documentary without some sympathy for Namath. He's been abandoned all his life: first by his father, then by his knees, then by his younger wife. (Even Pete Rozelle took a cheap shot at Namath with that Bachelors III power play in 1969. One day to sell? Insane). There's a scar on Joe's heart that's never going to completely heal.
I thought I had most of the broad points down on Namath (who won Super Bowl 3 about three weeks before I was born), but I learned a lot of stuff. I didn't realize he was initially bound for Maryland; I didn't know the details of the AFC Championship Game that preceded SB 3; and I certainly had no idea he was known for his fancy footwork at Alabama (much like Ken Stabler a few years later; similarly, he was a Stationary Snake in the pros). Growing up when I did, I knew Namath more for the Brady Bunch and the failed MNF opportunity and the Kolber incident (which is handled deftly - you can see Suzy's sympathy) than anything else. Well played, HBO.
When you hash it all out, Namath is the Mickey Mantle of his generation: handsome character from central casting; ideal backdrop of New York; family and alcohol issues; and a body that was, sadly, 20-40 years ahead of science. Imagine how good Mantle or Namath might have been if Dr. James Andrews existed back in his day. And the other thing about Mantle and Namath, I suppose, is that you had to see it live, you can't fully appreciate it as a backfill. There's no substitute for being there.
Don't sell Hernandez short simply because he's not as good as Gronkowski (and obviously he isn't). He has an excellent chance to be the X-factor in this game, lining up all over the formation. And just by using him as a diversion, you get the defense to give up information. No quarterback beats you before the snap quite like Tom Brady.
Brady, as mentioned earlier, is coming off one of his worst games of the year; I love his history in bounce-back spots. Manning's rating was low in the San Francisco win, but he's also riding eight TDs against one pick in the postseason. He's due to have one of his interceptions not dropped. And Belichick should have something tricky ready to go. You don't have to hold the Giants to 10 points here; you just need to stop them some of the time. And Belichick's resume is outstanding in rematch games, and revenge games (I'm talking about the Nov. 6 matchup - forget the 2008 result, the personnel has changed too much). Tom Coughlin would have preferred to play this game a week ago.
Look for the Pats to hurry up on offense, and slow the Giants down just enough on defense (maybe it's self-sabotage if the Giants run too much). New England 31, New York 27.