My First Post: Fortified With Stat Nerdery

<!--Introduction of myself, what I hope to write, then some statistical-mystical ramblings to follow-->

My name is CSStieber. I've trolled these MHR waters since the Buffalo Bills game last year (yeah, I froze my tukhus off at the game).

This was the MHR first post I read. hoosierteacher FTW

While I've only been ghosting around MHR for 5 months, I've been a dedicated sports-interwebs reader for going on 6 years. (I only write this in order to explain what will most assuredly influence my future posts) As a University of Michigan alum ('07), my home and hearth lay at MGoBlog.com, one of the premier college-sports blogs anywhere. EDSBS.com is where I also go for CFB fixes. As an inveterate NBA-nerdnik, FreeDarko.com has influenced me immeasurably in what I consider great sportswriting, where a fan tries, often in vain, to express in words the metaphysical connections between a his euphoria and it's relationship to his team's performances. I also dig FreeDarko's random-but-strangely-beautiful photos interspersed into the posts. So look for hints of all those sites in what I write.

Oh, yeah: Bronco bonafides: I graduated from a CoSprings high school in 2003, where the highlight in my sportsfollowing life up to that point was doing bumper cars with Eddie McCaffrey at Elitch's. I still think that Mike Anderson is underrated, both as a football player and as a hero for serving his country. I challenge any sportsfan with a soul to actually state that the Lambeau Leap is even in the same league as the Mile High Salute. Although a vegetarian, I would break my rules if it meant I could eat with John Elway at his steakhouse.

Okay, if anyone's still reading, follow me into the Mathemagical Land...

It took so long for me to find MHR because for years I thought that the NFL fan community was stupid. Yeah, I was a Bronco fan, but my impression of the league was that a math-minded internet nerd like myself was in the minority in NFL circles, while the collegiate sports community was much more likely to have people who thought like me. I mean, I LOVE the Barrel Man, but I just didn't see him enjoying regression analysis and efficiency statistics.

So when I finally took the plunge last fall, imagine how elated I was to find insightful and mature analysis of the professional game, for my own team no less!

In that vein, I'm gonna throw out some of the stats from both the Broncos and the Patriots and thoughts that came to mind as I looked at them. But first, I present to you CSStieber's Ironclad Rules of Predictive Statistical Analysis (*WARNING*: not really that Ironclad).

1. Ignore Gross Stats, For They Are Often A Result of Success Rather Than A Cause.  Yeah, not that earth-shattering, I know. In fact, the Divining the McDaniels Way Series covered this quite well. They even touched on some of the stats I like to use, but chose not to combine the inputs into my preferred conglomerate stats. The fact remains, however, that many times we as fans will see a gross-number stat (oftentimes presented unthinkingly by MSM TV analysts) and be impressed by its size. Not us at MHR, however. Say it with me: EFFICIENCY HAS THE HIGHEST CORRELATIVE RELATIONSHIP TO SUCCESS. Okay, that's a mouthful. I'll take just EFFICIENCY IS KING.

What do I mean by efficiency? Simple: if we were to run one single play facing an average opponent in an average gameplay scenario, what would be the standardized expected result? It's much easier to complete an 8 yard draw play on 3rd and 27. It's much easier to get an interception on a Hail Mary at the end of the half. We don't want those numbers to represent the proficiency an offense, defense, or individual. It doesn't matter if a quarterback throws for 3500 yards if he's throwing the ball 60 times a game. A giraffe could throw for that many yards if given enough opportunities. See: Drew Bledsoe. (I kid, I kid...)

Pertinent examples? Pertinent examples:

In 2003, the Notorious JMcD was a defensive backs assistant with the New England Patriots. Here's the vitals from that year for NE's defensive secondary (all stats from the ever-valuable Pro Football Reference):

328 completions on 618 attempts. 3232 yards given up, with 11 TD's and 29 INT's.

The 29 INT's led the league that year. Now, if we were to ignore Rule #1 of CIRPSA (above), we'd think, "Dang, that's impressive!" BUT WE WON'T. If you look, you'll see that the Patriots faced 618 pass attempts in 2003, which also led the league by an astounding 29 more attempts than second place. If we step away from the numbers and remember 2003, that would seem to ring true. The Patriots were 14-2 that year, so that means that many times during the season a team would fall behind to the Pats, which leads to more pass attempts (teams can rarely come from behind by using the time-consuming running game). Subsequently, the Patriots' gross stats would be inflated by the increased number of opportunities to pick off a pass.

Now, if we parse apart the numbers with some basic math, however, we can see just how effective the Pats' pass defense was on any single play. The nerd-standard numbers for efficiency, by my lights, are as follows: Interception Percentage (Int%), Adjusted Yards per Attempt (AY/A), and Adjusted Net Yards per Attempt (ANY/A).

Int% is derived from the simple equation of: INTERCEPTIONS/PASS ATTEMPTS. It's a good way to see how, on any given throw, what are the odds that the defense intercepts it. In this category, the 2003 Patriots were tied for second over all at 4.7% For comparison, that year's Broncos were dead last in the league at 1.8% That means that, on any given pass play, the Patriots were 250% more likely to intercept the ball. Yeeikes.

For AY/A, we can devise an equation that takes into account both passes completed and intercepted. For example, if there was a quarterback who equally threw touchdowns and interceptions-returned-for-a-touchdowns, he'd be a value neutral player on the field (statistically speaking, that is. Since nearly every QB brings some positive value to the field, one that brought 0 would be an abomination by comparison). In 2008, while the Broncos were 3rd and the Patriots were 12th in gross (boo!) yards, if you apply multipliers to the variables of touchdowns and interceptions, you can find better stats. I like PFR's formula of (PassYards + TD's*10 + INT's*45)/(PassAttempts), but we can quibble over the specifics if we want.  The AY/A for the Broncos in 2008 was 10th best, followed immediately by the Patriots in 11th, with only .2 yards between them. I think this case proves just how damaging Jay Cutler's 7 extra INT's were last year: they nearly completely negated the 900-yards and 4 more touchdowns worth of output from the Broncos' offense. Yikes, Part Deux.

(Oh, the Pats 2003 defense stifled opponents to a league-leading mere 3.3 AY/A. Sweet Fancy Moses.)

Finally, we can further tweak the AY/A by including the effect of sacks into the passing game, creating the ANY/A equation of (PassYards - SackYards + TDs*10 - INTs*45)/(PassAttempts + Sacks). This is, as far as I can tell, the single best stat to show that, when the offense drops back for a passing play, just what the average result will be. If a team defense has a great pass-rush or some coverage sacks, this statistic can count that into total defensive efficiency.

To briefly compare, again, the 2008 Broncos and Patriots: in ANY/A, they  were 8th and 13th respectively. (Another brief aside: the '08 Chargers were out of sight in terms of offensive efficiency, finishing well ahead of second place in both AY/A and ANY/A).

As for the 2003 Patriots' D (which was partly guided by Notorious JMcD): They tied for the league lead with the Baltimore Ravens, allowing a paltry 2.7 adjusted net yards per pass attempt. That means that the opposing offense could throw the ball on three straight plays and still face a 4 & 1.9! Yikes the Third.

Okay, thanks to you fearless few who've stuck around through this first post. I'm cutting it off here, but I promise to come back with more CIRPSA's, some less math-centric posts, and some interpretations I've made when looking over the performance of JMcD's past units.

I look forward to really diving into this disease we all call Bronco Madness, and hope to carry my weight round these here parts. Also, feel free to be as critical of my reasoning, research, body odor, whatever: any input is greatly appreciated.

Oaklandium delenda est!