Sabermetrics in hockey?

[PredsFan77*]

The Dodgers recently hired Oakland A's Assistant GM Paul DePodesta as their new GM. At 31, DePodesta is the third-youngest person to be hired as a big-league general manager. Theo Epstein was hired by the Boston Red Sox at age 28 in 2002, and Randy Smith was 29 when the San Diego Padres hired him in 1993. DePodesta, became the fifth or sixth GM that uses sabermetrics as factor when making roster moves. He also is noted as the numbers cruncher in the book Moneyball that showed the use of Sabermetrics as a viable strategy for baseball. My question is, do you all believe it is possible to come up with a similiar "formula" to calculate for hockey. Granted, the make ups of hockey and baseball are two completely different concepts, but are the statistics there to come up with these strategies?

If yes, do you think that they will ever come to the mainstream like in baseball,with owners who will buy into the sabermetrics reasoning? Will the NHL will see an influx of number crunchers from Ivy league schools becoming GMs before they are 40?

Click to read more...

 

[Dr Love]

I've thought about this long and hard... you can't. You can't do it in any sport other than baseball. In baseball, everything that happens is accounted for in a statistic. You can find worth of an all defense/no offense player like Ozzie Smith via statistic, yet you can't do the same for Bob Gainey. You can judge a player's defensive range via stats (if used properly, which barely anyone does) in baseball, but you can't account for the impact Keith Primeau has on a team's top line. You can't tell if someone is a "garbage goal" getter like John LeClair, but you can tell if someone is a slap singles hitter. There are just too many things unaccounted for in statistics in hockey to have anything close to Sabremetrics.

 

[MoS*]

Gary Bettman

 

[SuperNintendoChalmrs]

I buy into the Sabremetrics method as well.

Brian Campbell = Raffi Torres, Jason Smith, 1st rounder in 2004

catch the wave

 

[Burke's Evil Spirit]

What are sabermetrics?

 

[Dr Love]

What are sabermetrics?

Here's a 101. http://www.baseball1.com/bb-data/grabiner/manifesto.html

 

[PredsFan77*]

What are sabermetrics?

here's the gist of it...

http://www-math.bgsu.edu/~albert/papers/saber.html

Basically it's the statistical analysis of baseball.

 

[PecaFan]

It could be done, but this is a league that took 70 years before they started tracking save percentage, and still won't even publish scoring chances. Not to mention going backwards in the last few years, and hiding various stats.

 

[Pachoo]

No other sport lends itself as well to statistical analysis as much as baseball. But, I do think hockey could definitely be studied way better using that sort of analysis. For example, I love the PROD stat alot which calculates a players offensive production based on average playing time on ice. I like that as a gauge of prowess almost as much as overall points and goals. Of course, being surrounded by better linemates will help you in that area more than someone who doesn't, but it still is a good stat as a general barometer of play.

I also like Save Percentage more than Goals Against.

 

[Dr Love]

It could be done, but this is a league that took 70 years before they started tracking save percentage, and still won't even publish scoring chances. Not to mention going backwards in the last few years, and hiding various stats.

Scoring chances are subjective and not an offical stat, it's a stat made for TV as something to talk about. You can't do sabremetrics for strictly goalies, because their stats aren't indicative of the type of shots they face. Using stats wisely, you can tell what type of balls a player fields in baseball though.

 

[Dr Love]

No other sport lends itself as well to statistical analysis as much as baseball. But, I do think hockey could definitely be studied way better using that sort of analysis. For example, I love the PROD stat alot which calculates a players offensive production based on average playing time on ice. I like that as a gauge of prowess almost as much as overall points and goals. Of course, being surrounded by better linemates will help you in that area more than someone who doesn't, but it still is a good stat as a general barometer of play.

I also like Save Percentage more than Goals Against.

But PROD still doesn't tell the whole picture. You can't judge defensive worth with that, or if a player that is 23 is going to improve and by how much. But you make a fair point in that it is a good barometer, because it is, at the least if only because there isn't anything better.

 

[Lexicon Devil]

No way. Hockey is far too complex and creative a game to be statistically analyzed in the way a boring mechanical sport like baseball can be.

 

[Bruwinz37]

The Dodgers recently hired Oakland A's Assistant GM Paul DePodesta as their new GM. At 31, DePodesta is the third-youngest person to be hired as a big-league general manager. Theo Epstein was hired by the Boston Red Sox at age 28 in 2002, and Randy Smith was 29 when the San Diego Padres hired him in 1993. DePodesta, became the fifth or sixth GM that uses sabermetrics as factor when making roster moves. He also is noted as the numbers cruncher in the book Moneyball that showed the use of Sabermetrics as a viable strategy for baseball. My question is, do you all believe it is possible to come up with a similiar "formula" to calculate for hockey. Granted, the make ups of hockey and baseball are two completely different concepts, but are the statistics there to come up with these strategies?

If yes, do you think that they will ever come to the mainstream like in baseball,with owners who will buy into the sabermetrics reasoning? Will the NHL will see an influx of number crunchers from Ivy league schools becoming GMs before they are 40?

No, its impossible. Baseball is a series of individual events that can all be somehow formulated into numbers and percentages. You cant quantify forechecking, backchecking, fighting for pucks in the corners etc. Baseball you can quantify everything, but Hockey is more of a team sport. Baseball is very individualistic so it is easy to quantify a player's worth via sabermetrics.

 

[SmokeyClause]

No, its impossible. Baseball is a series of individual events that can all be somehow formulated into numbers and percentages. You cant quantify forechecking, backchecking, fighting for pucks in the corners etc. Baseball you can quantify everything, but Hockey is more of a team sport. Baseball is very individualistic so it is easy to quantify a player's worth via sabermetrics.

You are very close to nailing it. But, you guys are missing the most basic reason why baseball is the only sport that can effectively use Sabermetrics: Emotion. Baseball, sadly, can be played with little emotion. If a baseball player goes to the plate with extra effort, he's going home with a flyout (if he's lucky), if a pitcher goes to the mound with extra effort, he's going to be high in the zone. Baseball is a sport devoid of effort. Now, not in the truest since of the word, just in comparison to the other 3 major sports. Emotion is the one thing that cannot be calculated. Unlike other skills that can be reasonably quantified, this cannot. It's what seperates baseball from other sports. That combined with a deep statistical system and you have a wonderful tool to work with. In hockey, you can look at numbers all you want. You can even quantify how often someone is the first forward back, but you cannot quantify desire. You cannot quantify the effort someone exerts to play. It's just not measurable. And since emotion/effort is arguably the most key ingredient in hockey (it oftentimes outreaches skill as the determining factor in a game), it's virtually impossible to use a sabermetrical system to break hockey down.

 

[hbk]

I picked up a copy of Moneyball last summer and couldn't put it down. Fantastic read and this coming from a guy who's not a baseball fan by any means.

Actually, there are some lessons to be learned from Moneyball. Some fairly obvious lessons including:

1) Being successful without a large budget can be done if you do your homework and invest in player recruitment. Forget Sabremetrics. Invest in your scouting staff and system to maximize the ability of players to reach their potential. Too often teams cut these resources during time of fiscal constraint. Prime example is the Winnipeg Jets/Phoenix Coyotes who shared affiliation for a number of seasons and went off of the Central Scouting reports for the 1994 draft which yielded such gems as Deron Quint, Dorian Anneck, and Tavis Hansen. That year we dealt our first rounder for Dave Manson but regardless, this was probably the worst draft in team history. Would have been beneficial to have drafted a Jose Theodore or Patrick Elias. Hindsight is 20/20 and we are drafting teenagers which makes projection very difficult.

2) How to acquire more draft picks by working the system. Trading for free agents so you get the compensatory picks. Something for nothing is always good.

3) When trading always trade for the asset you want. Why deal anything for something you don't want or have a use for.

Great book.

 

[wint]

But, you guys are missing the most basic reason why baseball is the only sport that can effectively use Sabermetrics: Emotion.

IMO, the one thing that makes baseball so perfect for the sabermetrics method is that play is stopped and then started from the same exact setup countless times per game, with a limited number of variables being sufficient to completely describe each situation (i.e. men on base, inning, outs, count, who's hitting and pitching) and only one event taking place between stoppages of play (i.e. the result of the pitch). Also, there is only one player responsible for hitting on any given pitch.

Football is the next most ideal for sabermetrics (because it is played as a series of distinct downs). In basketball, you could separate the game into many possessions. But in hockey, it's harder to break the game up into pieces small enough for statistical analysis to be an effective way of representing what's going on.

There are far fewer restarts in hockey than in baseball (i.e. fewer faceoffs than pitches). It's often difficult to establish which team or player has possession of the puck at any given time, never mind make more difficult judgments that turn unique plays into quantitative representations. Plus, one of the chief purposes of sabermetrics is to establish the value of a player's value, to separate the individual's performance from that of his team. For many reasons, this is especially difficult in hockey.

Of course, some things in hockey are well-defined enough to fit the sabermetric mold - goals, shots on goal, minutes played, penalties and other infractions (like icing). And there can certainly be great improvements made to the current list of officially kept statistics (there must be a perfectly good stat for judging defensemen out there, but it certainly isn't plus-minus). In the end, though, hockey just isn't played in a way that makes it conducive to the sabermetric approach.

 

[Dr Love]

Football is the next most ideal for sabermetrics (because it is played as a series of distinct downs). In basketball, you could separate the game into many possessions. But in hockey, it's harder to break the game up into pieces small enough for statistical analysis to be an effective way of representing what's going on.

You can't do it for any sport though. There are no statistics that account for offensive line play, or a defensive tackle eating a block to allow a linebacker to make a tackle, or a running back making a play-saving block on a blitzer. Like I said before, baseball is the only sport you can do it in, because everything that happens on the field is accounted for in a statistic.

 

[Snakeeye]

The only reason sabermetrics exists is the structured and objective nature of baseball.

Nearly every stat in hockey is too subjective to lend itself to meaningful analysis. Some arenas inflate shot totals. Some hit totals. Also, there are too many factors that cannot be quantified in hockey. How do you punish the defenseman who let his man go, only to score? He gets the same -1 his teammates did, even if they did their jobs on that play.

There are more stats considered in sabermetric stats like EqA and VORP than exists in all of hockey.

 

[PecaFan]

Scoring chances are subjective and not an offical stat, it's a stat made for TV as something to talk about. You can't do sabremetrics for strictly goalies, because their stats aren't indicative of the type of shots they face. Using stats wisely, you can tell what type of balls a player fields in baseball though.

And hits and errors and fielding range of the shortstop etc aren't subjective? Of course they are. I see questionable scoring decisions all the time in baseball, because of just that, it's some guys subjective decision.

Yet baseball does just fine. There is no reason why scoring chances etc couldn't be an official stat.

 

[WhalerBoy]

I picked up a copy of Moneyball last summer and couldn't put it down. Fantastic read and this coming from a guy who's not a baseball fan by any means.

Actually, there are some lessons to be learned from Moneyball. Some fairly obvious lessons including:

1) Being successful without a large budget can be done if you do your homework and invest in player recruitment. Forget Sabremetrics. Invest in your scouting staff and system to maximize the ability of players to reach their potential.
2) How to acquire more draft picks by working the system. Trading for free agents so you get the compensatory picks. Something for nothing is always good.

3) When trading always trade for the asset you want. Why deal anything for something you don't want or have a use for.

Great book.

i agree!

I too read the book, couldnt put it down, and then started to see sports in a whole new light. I think if given enough time and statistics, hockey can be done this way too, to some degree. But really, for many of the reasons given above, especially the argument of emotion, baseball lends itself to this. Hockey does not.

But what got me hooked afterwards was could a hockey "fantasy" team be done this way, in otherwords, for hockey pools. Could you look back at players scoring only, and not worry about his defensive prowess or his PIM's or true hockey worth. Just does he score pts, will he break out this year.....if only i had the time, this would be interesting to dissect and follow. Try to win some more pools.

 

[Brain Hemorrhage]

They're beginning to use sabremetric type devices in the NFL...
footballoutsiders.com is the site i think. Not quite sabremetrics but its similar. budgoodesports.com is another NFL site that provides statistical analysis to NFL teams, including Bill Parcells (whatever team he is with when he's there, the unpredictable guy he is) who believes strongly in Bud Goode's formulas.