The GAA Battle

[CowbellGuy]

Didn't I answer this last week? Yes, it's counted as "Empty Net" when the goalie comes off for a delayed penalty. No, he doesn't get credit for playing goal when he's sitting at the bench.

[CowbellGuy]

[Q]billhoward Wrote:

 One can carry the math out to four, five, six decimal places to produce a goaltending number that's just slightly different from another ... but the margin of error is higher (for save percentage certainly) because of the vagaries of stat-keeping.

That's why three decimal places of precision for GAA is enough (1.xx) and two for save percentage (0.xx) where it's possible to be off by several shots a game on shots that may or may not have been saves.

Ditto for, say, points and assists per game. Does the ref always get that second assist right?[/q]

First off, we're NOT talking about SV%. There's no margin of error for GAA. The number of goals allowed is known and the time he's played is known.

Second, regarding SV%, would you say there's a margin of error on game scores? What if you don't think a goal should have counted, but the ref does? It's still a goal. Just the same, there's an official who counts shots on goal. Whatever he deems a shot on goal is a shot on goal. Those stats don't carry a margin of error; they're official. And since a shot on goal is 1.000000... shot on goal and not approximately one shot on goal, you can, and should, take the stats out to as many decimal places as needed to break a tie.

Whatever is on the official scoresheet, be it shots, goals, assists, or minutes played is all that matters and the only thing ever used to calculate stats, no matter what you happen to think about it.

[jeh25]

Bill is arguing false precision as far as I can tell...

Age is arguing that no such thing exists in sports because, as sports fans, we collectively agree to treat numbers veridically, noise and all.  Is there noise in the system? Certainly, but we all agree to pretend it doesn't exist.

I gotta go with Age on this one.

A hit is a hit, an at bat is an at bat and a goal is a goal. Do timekeepers shave seconds off the clock, affecting GAA? Certainly. Do homer score keepers inflate shots boosting SV%? Certainly. But for the purpose of records we all agree to pretend it doesn't exist.

Football has perhaps the best example of this convenient fiction with spotting the ball.  The ref spotting the ball is inherently inaccurate but we pretend it is an accurate event and then watch as they measure with the chains.

just my $.02

[DeltaOne81]

[Q]CowbellGuy Wrote:
First off, we're NOT talking about SV%. There's no margin of error for GAA. The number of goals allowed is known and the time he's played is known.

Second, regarding SV%, would you say there's a margin of error on game scores? What if you don't think a goal should have counted, but the ref does? It's still a goal. Just the same, there's an official who counts shots on goal. Whatever he deems a shot on goal is a shot on goal. Those stats don't carry a margin of error; they're official. And since a shot on goal is 1.000000... shot on goal and not approximately one shot on goal, you can, and should, take the stats out to as many decimal places as needed to break a tie.

Whatever is on the official scoresheet, be it shots, goals, assists, or minutes played is all that matters and the only thing ever used to calculate stats, no matter what you happen to think about it.[/q]
While there's not a margin or error on goals (or at least, not unless its a really contrived scenario), there is a margin of error on time played. Some score keepers take off for delayed penalty goalie pulls, some don't. There's of course a question of a few seconds on that even if you do. Is it when he leaves the net, when the other guy hits the ice, when he leaves the ice? Same goes for end of game empty net scenarios - do you take that out? Usually, but what's the timing exactly? A few seconds here or there different. So while the difference is certainly less than for SV%, there is still a practical margin of error.

That was my point for doing out the time difference that would make them equal (note I didn't do a goals difference).

Sorry to contradict you on several threads today, Age... it's nothing personal! :-D

[CowbellGuy]

Of course there's some error from the timekeepers, but whatever's on the official box score, for the purpose of keeping records, is the Word of God, and may be used with as much precision as anyone needs.

[KeithK]

[q]Of course there's some error from the timekeepers, but whatever's on the official box score, for the purpose of keeping records, is the Word of God, and may be used with as much precision as anyone needs.[/q]If what you're trying to do is decide on a qualitative level whether goalie A is better than goalie B it's silly to look at the third or further decimal place.  But if you want to say who had the better GAA or Sv% stat this year then the official stats are, as Age puts it, the Word of God.

[ugarte]

[Q]CowbellGuy Wrote:

 Of course there's some error from the timekeepers, but whatever's on the official box score, for the purpose of keeping records, is the Word of God, and may be used with as much precision as anyone needs.[/q]
Until, say, someone goes back and awards the 1910 batting title to Nap Lajoie.

[billhoward]

If we carry GAA out to 1.xxx instead of 1.xx, sooner or later we'll have a tie, there, too, that maybe is broken by going to 1.xxxx GAA.

As for save percentage, which others have remarked on as well - why is it listed as 94% not 94.2% on USCHO - one of the two component numbers is suspect. The number of goals allowed is pretty concrete (well, maybe Dartmouth was offside and it shouldn't have counted, but it did). The number of shots at the goalie is rife for potential error because a human is judging them, and from a lot farther away than the refs whose judgment is grist for a lot of postings here already.

The scorekeeper might favor the home team and see shots-just-wide (that are gloved) as saves for the home goalie, or he'll only notice two of the three shots hitting the goalie in a flurry in front of the net, or he'll count all three shots as shots on goal even though the second hit a defenseman not the goalie and the third was gloved wide of the net. Maybe the errors cancel themselves out, maybe not. Once the record book is signed, it's possible to carry the save percentage out to two, three, or four decimal places of precision, but that's not the same as two, three or four decimal places of accuracy. G.I.G.O.

Three and a half decades later, does it matter whether Ken Dryden saved 93.9 percent or 94 percent of the shots that the scorekeeper says Dryden faced? In some ways, it's probably better to remember 94 percent, because it's an accurate enough reflection on his career. Dave LeNeveu I believe was 93.8 or 94 percent on saves, take your pick, and if you go with the 94% number, it says these two remarkable goalies (or extraordinary and merely remarkable) goalies, playing a generation apart, stopped 94 of every 100 shots they faced. If those were in fact Dryden's and LeNeveneu's save percentages, is someone going to say LeNeveu was a better goalie? The difference is something like 2, 3 or 4 saves over the career: 938 vs. 939 saves out of every thousand shots faced.

A lot of this is a philosophical discussion and it's shaped by how you see the world or what your job is. I've had to explain with broad brush strokes to people who want to see the big picture and not have them caught up in minutiae, for instance buying one PC instead of a another because its Winstone score is 28.8 when someone else is 28.1. (Which all gets lost when the bleeping computer crashes every hour because the clock got pushed or some piece of software disagreed with XP and lost.) If I was the timekeeper for Nascar, then, yeah, I'd want the stopwatch to measure fractional seconds out to three places. And if I ever my eyes lasered, I think I'd like the surgeon to work out the cutting patters to about nineteen decimal places.

(True story. My colleague, five-six years ago, goes in for the eye surgery. She's patient No. 1 in the office at 9 a.m. She's waiting. She's nervous. She hears the doctor fire up the computer in the operating room. She hears the Windows 98 bootup chime. She knocks on the door: "Um, doctor, would it be okay if I was your *second* procedure today?")

[Robb]

[Q]Beeeej Wrote:

 I just sent a brief note to USCHO's Stats department.  I'm probably not only wasting my time, but obsessing a bit much.  

I understand that GAA is typically only *represented* in stats to the second decimal place.  But the fact that David McKee's GAA *rounds up* to 1.43 (1.427363078875) and Matti Kaltiainen's GAA *rounds down* to 1.43 (1.43060319585) tells me something's wrong with the way you're listing them on the stats page.  It's one thing to *list* them both has having a 1.43 GAA, that I understand.  I simply dispute whether it's accurate to list them as being *tied* by virtue of those "matching" 1.43 GAAs.

Beeeej[/q]
I'll sign your petition if you go along with mine asking that all scoring stats be on a "per game played" basis, the way goalie stats are done.

[Beeeej]

I've been shouting into the wind on that one for ages.  Gee, I wonder why there are no Cornellians on the Goals Scored leader board?  Could it be because we pay several fewer games?!  Blah.

Beeeej

[RichH]

[Q]Robb Wrote:
I'll sign your petition if you go along with mine asking that all scoring stats be on a "per game played" basis, the way goalie stats are done.
[/q]
Or, you could use collegehockeystats.com, where all but PPG, SHG, and GWG are on a per-game basis.
http://www.collegehockeystats.com/0405/national/d1m

[Beeeej]

[Q]RichH Wrote:Or, you could use collegehockeystats.com, where all but PPG, SHG, and GWG are on a per-game basis.[/q]

...and where Matti Kaltiainen is correctly listed as second, not tied for first, in GAA.

Beeeej

[Beeeej]

By the way, don't you love that our "clutch-and-grab" team is 49th out of 58 teams in penalty minutes per game for the season?

Beeeej

[ben03]

[Q]Beeeej Wrote:

 By the way, don't you love that our "clutch-and-grab" team is 49th out of 58 teams in penalty minutes per game for the season?

Beeeej[/q]
we just know how to get away with it ... duuhhhh, c'mon beeeej ::nut::

[ben03]

and if it weren't for mr. o'byrne we'd be even lower on that list