The Importance of Statistical Context

Let's start this thing with some hypothetical facts:

• TEAM A yields five goals per game.
• TEAM B scores five goals per game.

Your initial responses to each fact is likely the following:

• "TEAM A" has a hell of a defense.  A bunch of mercenaries out there, throttling folks for a bounty.
• "TEAM B" couldn't score at nickel night.

Each might be true, each might be false.  It all depends on context.  Let's add to the above facts with two more:

• TEAM A plays five defensive possessions per game.
• TEAM B plays five offensive possessions per game.

Well, TEAM A's defense isn't so hot; they're giving up a goal every time the opponent has the bean.  That's not good; in fact, it's double-plus ungood.  And TEAM B's offense -- that of a seemingly bookish nerd in a sea of wanting women -- is nothing but efficient, crushing it every time an opportunity presents itself.  That's the important of context.  That's why tempo-free statistics need relevance.  That's why I oftentimes have a problem with Quint Kessenich.

Now, Kessenich is a pretty good egg and he does a lot of his homework.  The problem with Kessenich (and it isn't just Quint; lots of guys make this error, including me) is that his statistical recitations are often without the appropriate context. 

In short, if you're going to tell me why a team is solid or Wagner-like, you need to show me the appropriate reasons that support the position.  Kessenich's latest Top 20 ballot and weekly analysis is riddled with absences of context to support his position.  Let's try and fix that.

19. Harvard - The Crimson resume includes wins over Canisius, Georgetown, Brown and Bellarmine. Jury is still out on the crew from Cambridge, but we do know that they are playing a fan-friendly version of lacrosse, taking 49.5 shots per game.

Our focus is on the bolded statement: What does taking a lot of shots have to do with fan-friendliness?  To me, a fan-friendly team is one that gets up and down the field, takes tons of shots, and buries a bunch of them as well.  I love 21-20 games, although those contests are long since dead (taking up residence next to short-shorts and knee-high socks.).  Here's the real deal with Harvard:

• The Crimson play 76.00 possessions per game, good for eighth fastest in the country.  So, Harvard's tempo (or style, if you will), is fan-friendly.  There's solid pace there.  You can't see that in a "per-game" metric; it's only see in a tempo-free, possession-based approach.
• Harvard is rifling 1.23 shots at the cage per possession, which is second most nationally. The problem with that, however, is that the Crimon's offensive effective shooting percentage is all of 41st-best in the country (26.49%).  To me, that isn't enjoyable.  That's like watching your grandmother work the remote control: She can eventually get the channel to ESPN, but it takes her forever to get the right digits.
• Harvard's offensive efficiency is moderate at 31.15 goals per 100 possessions (19th nationally).  I guess if you're a fan of Wagner you'd be enthralled with the Crimson's offense.
• As a contrast, Notre Dame takes the most shots per possession in the country (1.28).  The Irish's pace is 59th quickest in the land (56.00 possessions per game).  Unless you're Kevin Corrigan, there's no way that you're equating Notre Dame's style with fan-friendliness. So, if you're simply equating number of shots with fan-friendliness, you're not providing the whole story.  In fact, you're actually misleading the reader.

11. Johns Hopkins - The Jays have lost 10 of their last 11 to Top Ten teams, the latest an overtime defeat in the Dome. Hopkins owns the top scoring defense in the country, allowing just 5.4 goals per game with goalie Pierce Bassett leading all stoppers in save percentage (.683). JHU wins over Towson and Delaware were flattered this weekend. But until they beat a top team, they'll sit outside the upper echelon.

Again, we're focusing on the bolded statement.  I don't think Quint is making a poor point here (I'll explain in a second), I think that it's just incomplete:

• Johns Hopkins' defensive efficiency is second nationally at 18.53 goals per 100 possessions.  That's very, very good. As the Syracuse game showed everyone, though, something else is going into that equation: Pace.  The Blue Jays are only playing 65.43 possessions per game (40th in the country); they're also only seeing 29.71 defensive possessions per game (12th-fewest nationally).  So, while Johns Hopkins isn't yielding goals, it also is allowing opponents to have opportunities (possessions) to actually score. That, to me, is a bigger story than simply goals-for, goals-against; it is, rather, opportunity-for, opportunity-against when you're talking about the Blue Jays. 
• Bassett is stopping a high percentage of shots, but the Johns Hopkins defensive effort isn't totally his show. About 36% of the Blue Jays' defensive possessions end with save.  That's high -- only 11 schools see more goalie activity in terms of saves per defensive possession -- but Bassett is getting help from the guys in front of him.  You can see this in defensive assist rate (10th nationally at .1058) and defensive effective shooting percentage (19.78%, best in the land).  That adds up to more than Bassett stopping everything coming his way; that's a defense that is forcing shots eligible to be saved (e.g., bad shots).  Bassett has been terrific, but it's the entirety of the unit that is doing work.

5. Maryland - The Terps have a lone loss to Duke and are ready for the bulk of their 2011 schedule with upcoming games against North Carolina, Virginia, Navy and Hopkins. Maryland is No. 3 in the nation in goals per game with 13.2 and goalie Niko Amato sits at 65.2%. As a team, Maryland is shooting red-hot at 36%. Ryan Young has 16 points and Grant Catalino, the Big Cat, has found the net 15 times.

Kessenich appears to be implying the following: Maryland is winning because of its accurate offense and Amato's work in the cage.  To the bullet points!

• It's true: Maryland has a very good offense and it certainly has impacted the Terrapins start.  Maryland is ninth in adjusted offensive efficiency (32.69) and this efficiency is pushed, in part, by the Terps ability to can the bean (the team is sixth nationally in offensive effective shooting percentage at 35.07%).  To me, though, this isn't telling the entire story.  Something is missing.
• What's missing is how Maryland is getting on the board.  The Terps have been downright average on the extra-man this season (Maryland is converting on 26.32% of its extra-man opportunities, 37th nationally).  Given the low conversion percentage, it shouldn't be shocking that the Terps have virtually no reliance on the man-up unit to put goals on the board (only around seven percent of Maryland's goals are with the extra attacker; only two schools rely on the personnel imbalance less).  So, what we're left with knowing is that Maryland is getting tallies in even scenarios.  How does that offensive efficiency value look now knowing that Maryland is getting them in six-on-six? (I know, you're thinking that these are transition opportunities.  Well, that isn't really the case.  Maryland is 42nd nationally in pace (65.14 possessions per game.))
• Speaking of these even scenarios, the team's ability to shoot is great, but what's more important is how Maryland is getting those shots.  The Terps are first nationally in assist rate at .2422.  I would hope that Maryland would be in the top 10 in the country in shooting and offensive efficiency given that metric.  If you can get your hands free, set your feet, and get good looks at the cage, you better be converting.  And Maryland is. 
• Finally, while Maryland's offense has been very good, the Terps are also getting it done defensively.  Having to only play about 29 defensive possessions per game has helped (only five teams play fewer defensive possessions), Maryland has answered the bell: the team is seventh in adjusted defensive efficiency (20.96) and is sixth in defensive assist rate (.1)  Now, Kessenich gives some props to Niko Amato's work between the pipes (and Maryland does rely on the save a lot with 35.5% of their defensive possessions ending in save), but you also need to give some love to the rest of the defensive unit for it's work.