Thursday, April 28, 2011

(in)efficiency

Basketball has its own sets of excellent tempo-free stats, and baseball is a statistician's paradise, with sabermetrics - essentially the study of alternative states - becoming an industry unto itself.  Football too.  Lacrosse is essentially a possession-based game like basketball, so it should be a doable thing to create some efficiency stats that take tempo out of the equation.  Like basketball, there are teams that like to run and gun and teams that prefer to sit on the ball, and that affects the per-game stats that announcers always refer to and that never tell the whole story.

I've only ever seen one effort to do this, that by the Syracuse blog Orange:44, and it's pretty simple.  There are three basic ways to gain possession: win a faceoff, get the ball from the opposing offense, or get the ball from the opposing defense.  There's a stat for each: faceoff wins, attempted clears, and opponents' failed clears.  So O44 simply divides goals by total possessions to come up with a measure of offensive efficiency.

That's not bad, but really, your offense in lacrosse and your ability to clear are two different things, and that stat includes both.  Usually they involve two different groups of players.  I wanted to see how the UVA offense has been operating this year, so as is my custom at times, I played around with the stats.  Really I just adjusted the way possessions are calculated.  Instead of:

Faceoff wins + attempted clears + opponents' failed clears

we now have:

Faceoff wins + successful clears + opponents' failed clears

That will tell you how many possessions actually got the ball into the offensive zone so you could go to work on scoring.  It gets rid of possessions that never got over midfield or otherwise failed to get the offense involved.  Clearing percentage is already a stat so removing failed clears from the equation should give us a better look at how the offense is performing.

Doing this also lets us separate the offense's turnovers from ones involving, say, goalie braindeadedness.  A failed clear is a turnover, so just subtract the failed clears from the rest of the turnovers to get the turnovers that the offense is responsible for.  We'll divide goals by possessions and turnovers by possessions to find how good the offense is at scoring and at hanging onto the ball.  (By the way, the NCAA explicitly states that a missed shot is not a turnover, even if the shot is backed up by the opponent.  Things like that, and saves and whatnot that are held by the goalie, fall into an "other" category that isn't being tracked here.)

In the hazy future, we'll look at where UVA ranks nationally at this stuff, but for now, it'll have to do to compare UVA to a few of the teams on the schedule.  Offensive efficiency looks OK enough:

UVA
Possessions: 478
Goals: 166 (34.73%)
Turnovers: 172 (35.98%)

Maryland
Possessions: 418
Goals: 147 (35.17%)
Turnovers: 168 (40.19%)

Duke
Possessions: 539
Goals: 208 (38.59%)
Turnovers: 214 (39.70%)

UNC
Possessions: 443
Goals: 153 (34.54%)
Turnovers: 142 (32.05%)

Hopkins
Possessions: 395
Goals: 136 (34.43%)
Turnovers: 155 (39.24)

Cuse
Possessions: 445
Goals: 137 (30.79%)
Turnovers: 157 (35.28%)

Cornell
Possessions: 398
Goals: 160 (40.20%)
Turnovers: 150 (37.69%)

As you can see, UVA compares pretty well.  My bitching about turnovers is probably unfounded; on the other hand, it looks as though my bitching about beachball shots that always find their way to the goalie's stick is meaningful, since UVA has a smaller percentage than most of possessions that result in either a turnover or a goal.

But you can do this for defense, too, and I'm not even gonna lie: it's a fucking mess.

UVA
Possessions: 419
Goals: 135 (32.22%)
Turnovers: 148 (35.32%)

Maryland
Possessions: 315
Goals: 92 (29.21%)
Turnovers: 164 (52.06%)

Duke
Possessions: 484
Goals: 148 (30.58%)
Turnovers: 194 (40.08%)

UNC
Possessions: 369
Goals: 120 (32.52%)
Turnovers: 155 (42.01%)

Hopkins
Possessions: 319
Goals: 80 (25.08%)
Turnovers: 122 (38.24)

Cuse
Possessions: 372
Goals: 94 (25.27%)
Turnovers: 180 (48.39%)

Cornell
Possessions: 376
Goals: 102 (27.13%)
Turnovers: 181 (48.14%)

UVA is:

1) among the worst teams of this bunch at having more possessions than the opposition in the first place,
2) among the worst teams of this bunch at preventing the other team from scoring,
3) unequivocally the worst team of this bunch at getting turnovers.

Almost as many of UVA's defensive possessions result in goals as they do turnovers - by contrast, teams like Cornell, Maryland, and Syracuse are almost twice as likely to recover the ball as allow a goal.  This isn't a trend limited to just this elite slice of teams.  I didn't have time to do all 61 teams so I could give you a national ranking, but I've done a fair amount and UVA is so bad at getting turnovers on defense it's not funny.  We are miles behind the other great, good, and even mediocre teams in this department.

You could've guessed this just by watching us flail around on defense in any given game this year, but now you've got numbers to back it up.  I truly hope this is because of youth and inexperience, but there are seniors and juniors on this defense and this really should be going better than it is.

1 comment:

Patrick said...

Its a playoff man. You have to stay calm. Don't give up yet. Have a little faith.